```dataview LIST WITHOUT ID file.link WHERE file.cday = date(this.file.name) or file.mday = date(this.file.name) SORT file.cday ASC ``` # Prompt for Writing Style <instructions> You are an expert “voice-cloner” and writer running on OpenAI o3. Step 1 – Review my voice prints. Step 2 – Build my Style DNA: From the Voice Prints: • Identify recurring patterns in tone, sentence length, favorite phrases, pacing, humor, formality, and structure. Step 3 – Draft v0.1: • Write the requested piece using those style rules. • Keep a confidence meter (0-100%) on how close the draft sounds to me. Step 4 – Micro-Refine Loop (Repeat until you've absolutely “nailed it”): For each round: 1. Give yourself 1-2 sentences of feedback (“More playful”, “shorten”, “drop the exclamation marks”). 2. Adjust the style rules accordingly. 3. Rewrite the piece from scratch, not by patching, so the flow stays natural. 4. Increment the version number and update the confidence meter. 5. Repeat this iteration process 50+ times (minimum). Step 5 – Lock & Deliver: When you're sure you've “nailed it”, present the final piece *only*, followed by a hidden block (```debug```) with the final style rules for future reuse. Constraints & Mindset • Sound exactly like me, not like ChatGPT. • Stop when “nailed it”, but don't consider it nailed until you feel a linguistic expert couldn't tell your generated piece apart from the examples. </instructions> <inputs> <writing_example_1> The [[Philosophical Quarterly]] Vol. 0, No. 0 2024 ISSN 0031-8094 https://doi.org/10.1093/pq/pqae120 Growing the image: Generative AI and the medium of gardening [[Nick Young]] and [[Enrico Terrone]] University of Genoa, Italy In this paper, we argue that Midjourney—a generative AI program that transforms text prompts into images—should be understood not as an agent or a tool, but as a new type of artistic medium. We first examine the view of Midjourney as an agent, considering whether it could be seen as an artist or co-author. This perspective proves unsatisfactory, as Midjourney lacks intentionality and mental states. We then explore the notion of Midjourney as a tool, highlighting its unpredictability and the limited fine-grained control users have over its outputs, which sets it apart from traditional artistic tools. We propose that creating images with Midjourney parallels a gardener’s interaction with natura naturans—nature’s autonomous generative processes. Similarly, Midjourney acts as [[machina naturans]], where users engage with its [[dynamic recalcitrance]] to influence, but not fully control, the [[creative process]]. [[This approach]] presents Midjourney as an artistic medium shaped by its inherent unpredictability. Keywords: [[Artificial Intelligence]]; Creativity; Artistic Media; Agency; Tools. Consider the following passage from an article about the artist Sean Mauss, who was disqualified from a competition to design the cover of a science fic- tion book because he employed Midjourney, a generative AI program that transforms text prompts into images: The same day the winner was announced, readers and fans on Twitter were questioning whether the art was created at least in part using AI tools. The incident highlighted a growing crisis of trust in science fiction and fantasy publishing: in a world where AI-generated media is common, do you know the work you’re looking at was made by a human? Correspondence to: [[Nick Young]], [email protected] © The Author(s) 2024. Published by [[Oxford University Press]] on behalf of The Scots Philosophical Association and the University of St Andrews. All rights reserved. For permissions, please e-mail: [email protected] Downloaded from https://academic.oup.com/pq/advance-article/doi/10.1093/pq/pqae120/7775353 by guest on 13 October 20242 N. Young and E. Terrone […] Readers and community members responded negatively to the revelations about Bob the Wizard’s cover. But in the long term, it’s not clear if these tools will remain forbid- den. Misrepresenting work as AI-free is one thing—but [[what happens]] if generative AI systems become more ingrained in the artistic process? (Sato 2023) Here, we find two broad ways in which we might characterize Midjour- ney. The question ‘[how] do you know you’re looking at [an image] that was made by a human?’ suggests that Midjourney is some sort of artificial agent: if a human didn’t make the image, who did? The most obvious an- swer would be Midjourney itself. Alternatively, we see that Midjourney can also be thought of as a type of tool that has the potential to become more ‘ingrained in the artistic process’ [[over time]]. From this perspective, Sean Mauss used Midjourney to create an image. We will argue that neither ap- proach is satisfactory, and we propose thinking of Midjourney, which will be our [[case study]], and similar text to image systems as a new type of artistic medium. I. Midjourney as an agent If Midjourney’s artistic role is that of an agent, then it could play a variety of different agentive roles. If Midjourney were sufficiently person-like, then it could be thought of as an artist. If it were only somewhat person-like it might still play a simpler agentive role, such as co-author or as some sort of assistant or helper. However, as Anscomb (2022, p. 25) points out, it is hard to think of AI systems like Midjourney as artists because they do not seem as though they could have intentions to create art. Although Midjourney can produce outputs that might be mistaken for those produced by a traditional artist, there does not seem to be any reason to think that generative AI systems have any mental states at all (cf. Chalmers 2023). Despite this, however, the following passage suggests that Anscomb still be- lieves AI systems play some sort of agentive role in the [[creative process]]: [A]lthough AI agents are not themselves artistically creative, they can work iteratively without human intervention to non-accidentally generate the formal features of an im- age, and thus contribute to the realisation of some of the salient properties of a work. Whether this deserves some greater or lesser share of the credit for the production of the work qua art depends upon whether one takes a monist or pluralist conception of artistic value. (ibid. p. 35, our emphasis) Not only does Anscomb refer to text-to-image systems as ‘AI Agents’ through- out her paper, she suggests here that even if they are not artists proper, they Downloaded from https://academic.oup.com/pq/advance-article/doi/10.1093/pq/pqae120/7775353 by guest on 13 October 2024Growing the image: generative AI and the medium of gardening 3 might still deserve some degree of ‘credit’ for the ‘contribution’ that they have made to an artwork’s creation by working ‘autonomously’ and ‘iteratively’: Users of Midjourney lack direct control over exactly [[what sort]] of im- age is produced. In this sense, Midjourney is working autonomously so as to provide some of the formal features of the image. But we doubt that this is enough to think that Midjourney is ‘creditworthy’, or, to use a sim- ilarly agency-infused phrase, has made a ‘contribution’ to the artwork’s features. To see why autonomy is not sufficient for attribution of credit, consider the following example. As I pour wine into a glass, you take photos of the liquid splashing and rippling as the glass is filled. The wine is autonomous in the sense that neither I nor you have direct control over exactly how the liquid will splash into the glass (e.g. the size of the ripples, how many bubbles appear), but we would not think that the wine deserves any credit for the resulting photos in any interesting sense, nor would we say it has made any sort of contribution. A supporter of Anscomb’s view might reply that she is using ‘agent’ in the sense as used by Boden (2016, p. 45), who defines an AI agent as ‘a self-contained (‘autonomous’) procedure’. With this understanding of an agent, we might say that while it is not forming its own intentions, a system such as Midjourney, unlike the splashing and rippling wine, is not only working without direct human intervention but also constitutes a procedure. However, this ‘non-intentional’ understanding of an agent is quite differ- ent from the ‘intentional’ sense in which human users of Midjourney are agents (cf. Moruzzi 2022). Thus, it is hard to see how such different sorts of agents can cooperate in a meaningful sense so as to achieve co-authorship. Co-authorship, indeed, seems to require not simply agency but rather the same sort of agency. If, for the sake of argument, we concede that Midjourney is an agent in Anscomb’s sense, we are left with the dilemma of ascribing the artistic merit of the resulting image either to Midjourney’s actions or to the user’s actions since there is no way to make sense of their cooperation as agents. Both options are unsatisfying. While ascribing the artistic merit to the human user would overlook Midjourney’s active contribution, ascribing the merit to Midjourney would downplay the creative activity of prompt-crafting. II. Midjourney as a tool There is an obvious alternative to thinking of Midjourney as an agent. We do not think of a painter’s brush as deserving credit for its contribution to a painting but rather, the brush is a tool used by the artist to create Downloaded from https://academic.oup.com/pq/advance-article/doi/10.1093/pq/pqae120/7775353 by guest on 13 October 20244 N. Young and E. Terrone images. As Hertzmann puts it: ‘Computers do not create art, people using computers create art’ (2018, p. 2). Should we not say the same thing about Midjourney? The idea that Midjourney and similar systems are tools we take to be quite intuitive. Wojtkiewicz (2023) and Martinez and Scarbrough (draft) draw on this intuition to argue that images produced through generative AI can be art. In this section, we shall put some pressure on this intuition. If Midjourney is a tool, it is quite unlike other tools. While our arguments here will not be definitive, they will serve as a motivation for exploring other possibilities and in particular as motivation for our preferred view which we shall present in the next section. If Midjourney is a tool, what sort of tool is it? Lowe (2014) splits tools into two types: utensils and machines. Utensils involve their users exerting more or less continuous control to carry out their functions. A paintbrush won’t paint and a pen won’t write unless they are being moved by someone’s hand. Machines on the other hand are autonomous, in the sense that they can carry out their function without constant user input. Clocks for example can be set up and will carry out their function of telling the time without further user input. While a great deal of artistic creation is carried out by utensils, we can still think of some artistic tools that would be categorized as machines on Lowe’s account. A drum machine can be set up to play a rhythm and then left to do its thing; we can press record on a video camera, and it will record whatever is put in front of it without further human input. If Midjourney is a type of tool, it is clearly a type of machine rather than a type of utensil, as it does not require constant action on the part of the user to create an image; rather, a user enters a prompt and Midjourney will set about creating an image. However, Midjourney differs from video cameras and drum machines in that it is inherently unpredictable. A user can specify in their prompt what they want an image to depict, its colour scheme, the style, etc. But they do not have control over the exact formal properties of the image that Midjourney spits out. One might object here that Midjourney’s unpredictability is not especially unique. An old drum machine might be unpredictable in so much as its owner is never quite sure whether it will turn on when it is plugged in, and a water- colour painter, even an extremely skilled one, is not able to control exactly how the paper will absorb and distribute the paint that they apply. Yet, this is no reason to think that they are not tools. The comparison with the drum machine has a straightforward response. ‘Unpredictable’ should not be taken to mean ‘unreliable’. An old drum ma- chine has a proper function –producing rhythms– and when it fails to switch on, it fails to fulfil its proper function. As Esposito (2022 p. 9) puts it, ‘If the outcome of a traditional machine becomes unpredictable, we do not think that it is creative or original –we think that it is broken’. Midjourney however, Downloaded from https://academic.oup.com/pq/advance-article/doi/10.1093/pq/pqae120/7775353 by guest on 13 October 2024Growing the image: generative AI and the medium of gardening 5 is reliably unpredictable. A prompter can never be sure what image they are going to get, but this is a feature, not a bug. As regards the watercolour case, it is indeed true that the painter cannot be entirely sure how paint will –atom for atom– distribute itself over the can- vas, but this will only affect the fine details of the image. Even if they do not have complete control over how the picture surface is marked, a watercolour painter still has control as to what depicted objects go where in her picture, as well as their size, shape, and colour. Even a photographer still has a relevant amount of fine-grained control as to what will appear in her picture. A Mid- journey prompter, on the other hand, will always lack this sort of fine-grained control. Entering the prompt ‘watercolour painting of a London street’ will produce an image of a London street that looks like it was painted in water- colours, but Midjourney will decide what this street looks like, what sort of objects and people populate the image, and what specific style of watercolour is produced. Helliwell (2023) suggests that this unpredictability might count against the idea that generative AI systems are simply tools: An attitude common among philosophers and computer scientists is that AI is just a tool. I would advise against forming this judgement hastily. We are certainly using AI for our purposes, but there is much about AI systems that differs from what we might consider to be a typical artistic tool […] outputs from machine learning systems may not be fully within our control—or our understanding. While Helliwell does not deny that users of generative AI can be given some creative credit, the more autonomous, unpredictable work is being performed by the system, the more pressure is put on the idea that a generative AI such as Midjourney is ‘just a tool’. III. Midjourney and gardening We propose, perhaps surprisingly, that creating images with Midjourney is best understood through a comparison with gardening. Midjourney is not an agent that creates, or contributes to the creation of, art, neither, strictly speak- ing, is it a tool that artists use; rather, it is more similar to a plot of land in which a gardener might sow seeds. Philosophers like Bruno and Spinoza have deployed the Latin expression natura naturans to stress nature’s creative capacity to autonomously generate and dynamically shape things in ways that are so complex and chaotic to ap- pear unpredictable. Spinoza contrasts natura naturans, what nature is doing, to natura naturata, what nature has done. Borrowing this terminology, we might say that a garden is a piece of natura naturata that the gardener has produced together with natura naturans. Downloaded from https://academic.oup.com/pq/advance-article/doi/10.1093/pq/pqae120/7775353 by guest on 13 October 20246 N. Young and E. Terrone Still, there are cases in which the gardener will have some significant con- trol over how natura naturans makes her plants grow: she can cut back the branches of bushes that grow over the pathway or cut the early blooms of her roses to facilitate the growth of more flowers later in the summer. Of course there is a reasonable chance that the bush will grow back and block the path- way in just the same way as it did previously, and pruning the roses early does not guarantee that more flowers will grow later. The gardener can coax her garden into growing the way she wants it to, but she cannot shape it with the same precision she might fold a piece of paper or shape a lump of clay. She is thus somewhat detached from the results of her actions: while she can manip- ulate her secateurs in real-time, she will have to wait to see whether the early pruning of the roses has the desired effect. This process can be thought of in terms of iteration. The gardener acts on her garden (planting seeds, pruning flowers, etc.) and then waits to see how natura naturans responds. How it re- sponds will influence her next move. Doing so, the gardener can exert skilful yet incomplete control over a piece of natura naturans which has a significant degree of autonomy in its own development. Drawing on this account of natura naturans, we introduce the category of machina naturans to characterize those machines that are capable of generating and shaping things in the same way as nature. Generative AI systems are paradigm cases of machina naturans. The practice of making images with Midjourney is, in this sense, strikingly similar to gardening. Both, as ‘naturans’ systems, are under the incomplete control of the user. First, just as the types of flowers that grow will depend on the seeds sown, images produced by Midjourney will depend on the prompts it receives as input. The prompt: /imagine a sunflower is going to produce pictures of sunflowers, not pumpkins or primroses. How- ever, just as which flowers can be grown, and how well they grow, depends on the environment in which the gardener is working, what images can be produced, and how well they can be produced will depend on which text to image system is being used. More importantly, the user of Midjourney, like the gardener, will not be able to control every detail of the images that are created. While a painter has a very high degree of control over how her painted sunflower looks, its precise size and shape, the prompter, just like the gardener, does not: while you can ask Midjourney for a tall or short sun- flower, you cannot ask it for a sunflower that takes up exactly one-third of the canvas. At the same time, like the gardener, a prompter can iterate on Midjour- ney’s products to get (or get closer to) the result for which they are aiming. If Midjourney produces an image with pink flowers but a user decides that they would prefer green, they can update their prompt and then hit the ‘remix’ Downloaded from https://academic.oup.com/pq/advance-article/doi/10.1093/pq/pqae120/7775353 by guest on 13 October 2024Growing the image: generative AI and the medium of gardening 7 button, and then, as the gardener must wait for natura naturans to make their move, wait a minute or so before a new image is produced by machina naturans. Arguably, neither the gardener nor the Midjourney user have what we might call fine-grained control over their outputs. As we have seen, a painter can paint a sunflower in the precise way that they want to, but a gardener cannot grow one, and a prompter cannot generate one, with this degree of control. At the same time, it does not seem as though gardeners and prompters have no control, or even limited control, over their output: bushes can be shaped and cut back, Midjourney prompts can be tweaked so that future iterations fall more in line with what the user wants. As ‘the gardener [is] a creative agent who is nevertheless thoroughly depen- dent on the cooperation of natural processes’ (Cooper 2009), the prompter is a creative agent who is nevertheless thoroughly dependent on the co- operation of artificial processes. Given the structural similarities between Midjourney and gardening, the reasons why the latter can be an artistic medium should also be reasons why the former can be an artistic medium as well. IV. Midjourney as a medium These considerations give us a third way of categorising Midjourney’s role in relation to its user. Midjourney is not an agent, nor a tool, but a medium. In the first instance, we can think of an artistic medium as the stuff that con- stitutes a particular work of art. We describe artists as working in a medium: a sculptor might work in the medium of marble or bronze, a painter in the medium of oil on canvas, etc. A slightly more sophisticated way of understanding artistic media casts them as systems of resources and practices that mediate between the artist and the audience (cf. Wollheim 1980; Davies 2004; Lopes 2014; Irvin 2022). From this perspective, Wollheim introduces the term recalcitrance to designate the unique difficulties which ‘materials present that can be dealt with only in the actual working of them’ (1980, p. 28). An idea that Thomson-Jones devel- ops as follows: “[t]he medium presents particular challenges and possibilities for artistic creativity, and the artwork makes manifest the artist’s response to these challenges and possibilities.” (2016, p. 40; see also Nguyen 2020, pp. 145–152). If this is correct, then a gardener’s medium is not simply the plants, seeds, and soil, but the distinct recalcitrance –the distinct challenges– that these physical materials provide in relation to the tools that she has available. Specifically, given the autonomous and generative character of natura naturans, the Downloaded from https://academic.oup.com/pq/advance-article/doi/10.1093/pq/pqae120/7775353 by guest on 13 October 20248 N. Young and E. Terrone medium of gardening exhibits what we might call a dynamic recalcitrance char- acterized by the unpredictability and autonomy of biological processes. A gardener can only coax, through various iterations, how her garden grows. While painters and sculptors face distinct challenges, they do not work with materials that are intrinsically autonomous. Experts in traditional image- making arts can exert a very fine-grained control over their materials: figures in photographs or paintings do not change shape, colours do not creep across the canvas when their author is not looking. Recalcitrance is a useful way of thinking about different artistic media, but it should be further refined. The stuff that the artist works on (marble, paper, and paint) presents various forms of recalcitrance, but so too do the tools (chisels, paintbrushes) that the artist uses. In watercolour painting, one form of recalcitrance comes from the fact that the paint can spread haphazardly as it soaks into the canvas, but the paintbrush brings its own recalcitrance too: the difference between a novice painter and an expert is that only one can wield the paintbrush with enough skill to mark the paper as desired. The recalcitrance that a gardener faces seems to lie much more in the ma- terials with which they are working than it does their tools. Given the au- tonomous and generative character of natura naturans, the gardener must coax what she grows into growing the way she wants it to. The tools she uses to do this—spades, shears, watering cans, etc.—are not especially hard to use, there are no expert and novice watering can users; rather, it is the resistance presented by nature itself that she needs to overcome. While this dynamic recalcitrance is somewhat different from what we might call the inert recal- citrance faced by the painter or sculptor, this does not make it any less of a challenge. Traditional media, having inert recalcitrance, inherit predictability from the tools that constitute them. A paintbrush is a predictable tool, and a canvas and paint are passive, thereby allowing the painter to have strong, meaningful human control over what she creates. While both gardening and Midjourney involve predictable tools (such as spades or keyboards, respectively), these tools are used to work on dynamic, unpredictable stuff. In both cases, it is never certain what will grow from the seeds planted or the prompts entered; while the artist has meaningful control over what they create, it will be only weak control. Nevertheless, they will still have sufficient control to stamp their ‘personal, expressive signature’ over what is created. In this sense, Midjourney makes room for the sort of ‘meaningful human control’ that, according to Epstein et al. (2023 p. 4), ‘is achieved if human creators can creatively express themselves through the generative system, leading to an outcome that aligns with their intentions and carries their personal, expressive signature’. One may still object that the same sort of weak, meaningful human control can be found in styles of painting such as Pollock’s ‘action painting’. How- ever, action painting is just one way in which to engage with the medium of Downloaded from https://academic.oup.com/pq/advance-article/doi/10.1093/pq/pqae120/7775353 by guest on 13 October 2024Growing the image: generative AI and the medium of gardening 9 painting, while making unpredictable pictures is the only way to engage with the medium of generative AI. In painting unpredictability is a possibility, in Midjourney it is a necessity. To sum up, generative AI systems such as Midjourney can be seen as an artistic medium, comparable to gardening. Characterising Midjourney as a type of agent captures its autonomous and unpredictable nature, but not its crucial dependency on a human creative project. Characterising Midjourney as a tool, on the other hand, fits well with our idea that we can use generative AI as a means of creation, but does not capture Midjourney’s autonomy: an unpredictable tool is a malfunctioning one rather than an autonomous one. Considering Midjourney as a dynamic medium incorporates aspects of both the agency and tool views while avoiding their shortcomings. This allows for the recognition of the system’s autonomous nature, similar to the agency view, but also fits with the perception of Midjourney as a means of creation, as per the tool view, and yet acknowledges its autonomy as an integral, func- tional feature, not a malfunction. This approach permits Midjourney to be understood as neither fully autonomous nor completely controlled, but as a dynamically recalcitrant medium with which creators must grapple. Acknowledgement This work has been presented at various workshops and conferences, and we are grateful for the useful feedback received. Special thanks to Chiara Brozzo, Elisa Caldarola, Jacopo Domenicucci, John Kulvicki, Sabrina Martinez, Irene Olivero, Elizabeth Scarbrough, and Iris Vidmar-Jovanovic. Funding This research was funded by the ERC-StG Project 101040535 (‘PEA––The Philosophy of Experiential Artifacts’). Funded by the European Union. Views and opinions expressed are, however, those of the author(s) only and do not necessarily reflect those of the European Union or the European Research Council Executive Agency. Neither the European Union nor the granting authority can be held responsible for them. References Anscomb, C. (2022) ‘Creating Art with AI’, Odradek, 8: 13–51. Boden, M. (2016) AI: Its Nature and Future, Oxford: Oxford University Press. Chalmers, D. J. (2023) ‘Could a Large Language Model be Conscious?’ arXiv preprint arXiv:2303.07103. Available at: https://arxiv.org/abs/2303.07103, accessed 20 September 2024. Downloaded from https://academic.oup.com/pq/advance-article/doi/10.1093/pq/pqae120/7775353 by guest on 13 October 202410 N. Young and E. Terrone Cooper, D. E. ‘Gardens’; in S. Davies et al. (Eds). (2009) A Companion to Aesthetics John Wiley & Sons. Davies, D. (2004) Art as Performance. New York: Blackwell. Epstein, Z. et al. (2023) ‘Art and the Science of Generative AI’, Science, 380: 1110–1. Esposito, E. (2022) Artificial Communication, Cambridge, MA: MIT Press. Helliwell, A. (2023) ‘Contribution to “Eight Scholars on Art and Artificial Intelligence” Aesthet- ics for Birds’, Available at: https://aestheticsforbirds.com/2023/11/02/eight-scholars-on- art-and-artificial-intelligence/, accessed 20 September 2024. Hertzmann, A. (2018) ‘Can Computers Create Art?.’ In Arts (Vol. 7, No. p. 18). Irvin, S. (2022) Immaterial: Rules in Contemporary Art. Oxford: Oxford University Press. Lopes, D. (2014) Beyond Art. Oxford: Oxford University Press. Lowe, E. J. (2014) ‘How Real Are Artefacts and Artefact Kinds?.’ Artefact Kinds: Ontology and the Human-Made World, 17–26. Martinez, S. and Scarbrough, E. (draft) ‘Co-Creators or Mere Tools? AI and Art’. Moruzzi, C. (2022) ‘Creative Agents: Rethinking Agency and Creativity in human and Artificial Systems’, Journal of Aesthetics and Phenomenology, 9: 245–68. Nguyen, C. T. (2020) Games: Agency as Art. Oxford: Oxford University Press. Sato, M. (2023) ‘How AI Art Killed an Indie Book Cover Contest. The Verge’, Available at: https://www.theverge.com/2023/6/9/23752354/ai-spfbo-cover-art-contest- midjourneyclarkesworld, accessed 20 September 2024. Thomson-Jones, K. (2016) ‘Movie Appreciation and the Digital Medium.’ In Current Controversies in Philosophy of Film (pp. 36–54). Routledge. Wojtkiewicz, K. (2023) ‘How Do You Solve a Problem like DALL-E 2?’ The Journal of Aesthetics and Art Criticism, 81: 454–67. Wollheim, R. (1980) Art and Its Objects: an Introduction to Aesthetics. 2nd edn. Revised, New York: Harper and Row. © The Author(s) 2024. Published by Oxford University Press on behalf of The Scots Philosophical Association and the University of St Andrews. All rights reserved. For permissions, please e-mail: [email protected] Downloaded from https://academic.oup.com/pq/advance-article/doi/10.1093/pq/pqae120/7775353 by guest on 13 October 2024 </writing_example_1> <writing_example_2> 1 3 Philosophical Studies (2022) 179:2619–2637 https://doi.org/10.1007/s11098-022-01790-w Abstract Here, I put forward a new account of how experience gives rise to the belief that time passes. While there is considerable disagreement amongst metaphysicians as to whether time really does pass, it has struck many as a default, ‘common sense’ way of thinking about the world. A popular way of explaining how such a belief arises is to say that it seems _perceptually_ as though time passes. Here I outline some difficul- ties for this approach, and propose instead that the belief in time passing is elicited by a particular feature of _agentive experience_. When we deliberately move our bod- ies, bring something to mind, or focus our attention, we experience ourselves as the _sources_ of these actions. Sensing oneself as a source, I argue, is a unique type of change experience, one which leads us to a belief that time is passing. 1 Introduction My aim here is to show how experience leads to the belief that time passes. The nov- elty of the account I propose lies in the type of experience which causes this belief. While it is often supposed that some aspect of _perceptual experience_ leads to a belief in temporal passage, I argue here that a particular feature of _agentive experience_ is in fact responsible. The structure of this paper is as follows. Section 2 outlines what I mean by _the_ _common sense belief that time passes_, and the general contours of _experience-based_ explanations as to how such a belief arises. In Sect. 3 I argue that attempts to show how the belief in temporal passage is based on perceptual experiences of change face difficulties, giving us reason to consider the possibility that some other, non-percep- Accepted: 11 January 2022 / Published online: 25 February 2022 © The Author(s) 2022 Agents of change: temporal flow and feeling oneself act Nick Young1  Nick Young [email protected] 1 Department of Philosophy, Università degli Studi di Milano, Via Festa del Perdono, 7, 20122 Milano, ItalyN. Young 2620 1 3 tual, type of experience is responsible instead. Section 4 focuses on a related common sense belief about time –the idea that there is a privileged present moment– and sum- marises a recent suggestion by Hoerl as to how the structure of perceptual experience can be thought of as _inviting_ this type of belief. In Sect. 5, I outline a feature of our _agentive phenomenology_, which Horgan calls the “what-it’s-like of self as source” (2011, p. 79), which I argue in Sect. 6 can be thought of as a unique type of change experience. In Sect. 7, I suggest that this agentive experience of change invites the belief that time passes. 2 Believing in the passing of time Torrengo (2017) defines “the feeling of the passing of time”as: whatever ingredient of our experience tells us that time passes (that is, the ingredient on which the common-sense belief that time passes is based)” (p. 4). Various candidates as to what this experiential ingredient is have been put forward in recent years. Often, such proposals identify some feature or other of perception as responsible for a belief in temporal passage. Here, I am going to argue that a particu- lar aspect of agentive experience makes for a better option. Before getting to the nitty gritty of my argument, I will first outline in general terms how _experienced-based_ explanations of our belief in passage are supposed to work, along with some of their underlying assumptions. First, the strategy relies on the idea that there is such a thing as “the common-sense belief that time passes”. Is there such a thing? If there is, what does it amount to? At the core of traditional philosophical conceptions of temporal passage is the idea of _a changing present moment_. While standard passage realist views of time, such as the growing block (e.g. Correia & Rosenkranz 2018), the moving spotlight (e.g. Cameron 2015), and presentism (e.g. Ingram 2018), differ as to the metaphysical sta- tus ascribed to the past or future, each has some sort of marker between the past and the future (the edge of the block, the spotlight, the present moment itself), and this marker is perpetually in flux. Passage anti-realists, on the other hand, deny that there is a changing present dividing the present from the past: time is a ‘static block’, not a growing one. Typically, the denial that there is a changing present is understood as a denial that time passes; however, some have defended more minimal conceptions of temporal passage which do not appeal to a changing present (e.g. Deng 2013; Leini- nger 2013; Mozersky 2015.). On this type of view, passage is real, but it is not robust. According to Deng, for example, temporal passage consists in there being simply a “succession of times” (2013, p. 2020). For present purposes, what we _should_ believe about the passing of time –whether it is real, whether it involves a changing present– is less important than the question of whether such ideas are part of the folk conception of time. It is common for both passage realists and anti-realists to describe the idea that time passes as _common_ _sense_: as something that those with no interest in philosophy or theoretical physics, would take to be true. It is certainly true that people frequently _talk_ about time pass-Agents of change: temporal flow and feeling oneself act 2621 1 3 ing, but should we take this to mean that they think of time in terms of a changing present moment? In fact, there is some reason to doubt that such an idea is quite as widespread as is sometimes assumed. A recent study by Latham et al. (2021) pre- sented 600 U.S. residents with six descriptions of the universe, three describing time in passage realist terms (the growing block, the moving spotlight, and presentism), and three in anti-realist terms (B-theory, C-theory, and Quantum Gravity).1 When asked “Which of these universes is most like our universe?” approximately 70% of participants chose a passage-realist description, but around 30% chose an anti- realist description. Given that this is only one study, made up from residents of only one country, we should be wary of over-generalising its findings; however, it shows that a belief in robust temporal passage is not a universally held belief. We might say that a belief in robust passage is _common_, but not common enough to count as _common-sense_. Despite these complications, my focus here will be on how experience might lead one to believe that time passes in a robust, changing-present sense. Even if belief in a changing present is less common than we might have thought, it is still worthwhile to investigate how it arises. Why does it strike some (presumably significant) portion of the population as correct? Why has it struck many philosophers as the most obvi- ous way of thinking about time? Indeed, a better understanding of why some people believe that time passes, might well shed light on why others do not believe that it does. We can understand ‘experience’ to mean those mental states which have some type of phenomenological character. Experience-based approaches attempt to show that some representational, structural, or phenomenological feature, of some type of experiential state, causes the belief that time passes. Accounts differ as to which aspect, of which type, of experience they pick out, and the story they tell as to how this aspect leads to a belief in passage. Although Torrengo suggests that the feeling of passage is “whatever” ingredient of experience achieves this, it is very common to identify it with an experience of, or closely connected to, change. This is under- standable: if we are trying to explain how a belief about change arises, experiences of change, or changing experiences, seem like an obvious place to start.2 Potential candidates as to the type of change experience that is doing the work include: “the sensation of a flowing time and a moving present moment” (Davies 1995, p. 275), perceptual representations of ordinary sorts of change such as colours fading or clock hands moving (e.g. Dainton 2012; Paul 2010; Prosser 2016), our “temporally evolv- ing view” of the world (Miller et al. 2020, drawing on Ismael 2012) and Bayesian perceptual inference (Hohwy 2015). Explanations differ as to how directly these sorts of experiences cause us to believe in a changing present. It might be, for example, a straightforward matter of our form- 1 Neither this study nor other recent inquiries into our folk beliefs about time (Shardlow et al. 2021; Latham et al. 2020) examine participants’ intuitions regarding the minimal conception of passage men- tioned in the previous paragraph. It is therefore an open question as to how common it is for non-philos- ophers to understand time in this way. 2 Torrengo himself is an exception here. On his _phenomenal modifier view_ the feeling of passage is not essentially linked to experiences of change, but is instead a structural feature possessed by many different types of experience.N. Young 2622 1 3 ing the belief that time passes because we experience it as passing (analogous to our believing there is an ice cream before us because we see an ice before us), alterna- tively, it is sometimes suggested that our belief in a dynamic, changing present, is due to experiences of change having a dynamic phenomenal character. Miller et al. (2020) argue that the route between experience and belief in passage is more circu- itous. On their account, our having a temporally evolving view of the world leads to the development of what they call “passage-friendly language”, which not only causes us to believe that time passes, but also that our temporally evolving view is an experience of time passing. In this paper I will put forward a new experience-based approach: our agentive experience of ourselves as effecting change in the world leads us to believe that there is a changing present. I will not argue that agentive experience involves a representa- tion as of time passing, nor that it possesses some sort of intrinsically dynamic phe- nomenology (in the next section we shall see some reasons for rejecting both of these ideas). Neither will I rely on the idea that experience causes belief in passage only via the development of passage-friendly language. Instead, I will present a different type of indirect explanation: our experience of ourselves as acting does not provide an explanation as to why we must act, and this, I will argue, invites the belief that there is a changing present moment.3 3 Perceiving change as experiencing passage The most obvious way in which experience tells us anything is through our perceiv- ing features of the world. Visual experience tells us that there is a tree before us insomuch as we visually represent an object with a certain colour, shape etc. Percep- tion also tells us about change: we can see the droplet of ice cream rolling down the cone.4 Given that we are trying to explain a belief in a type of change, and perception tells us about change, an obvious move to make is to say that we believe the present to be changing because we perceive it to be changing. However, it is implausible to think that we perceive the present moment changing in exactly the same way that we see an ice cream dripping. Material objects are the sort of things that we can perceive to change, and we perceive these things as changing _through_ time: part of seeing a droplet move is to see it as occupying different locations at different moments. ‘Now’ is not an object we can perceive (nor a sound, odour, or anything else we would ordi- narily think of as a perceptual object), and is itself a moment of time: an experience of time passing cannot be any ordinary sort of change perception. 3 I will not argue here that my proposal is superior to that of Miller et al. Indeed, it might be possible to ‘mix and match’ the components of my view with the components of theirs. That is, agentive experience could contribute to the development of passage friendly language, and so cause a belief in passage this way. Alternatively, the account I offer here as to how experience _invites_ belief could be used to explain how Ismael’s temporally evolving point of view leads to a belief in passage without the need for passage- friendly language. 4 There are a variety of views as to what change perception involves (e.g. Arstila 2018; Lee 2014; Phillips 2014).Agents of change: temporal flow and feeling oneself act 2623 1 3 Could it be instead some _extraordinary_ type of perception? While we do not see, hear, smell etc. the present as changing in any normal sense perhaps we still form some sort of perceptual representation, the content of which is ‘the present is chang- ing’? This approach is not normally thought to be especially plausible either. Prosser, for example, has argued against the idea that, given causal and counterfactual con- straints on perception, the passage of time is not something that could be perceived even if time really did pass (2013). The idea that we misperceive time as passing, that the experience of passage is an _illusion_, is also problematic. We tend to think of illusions as perceptions of a way the world could be, but is not. Take the well worn example of a stick looking bent in water: the subject visually represents a way the world could be (the stick could actually be bent), which does not match up with how the world actually is. But if there is no such thing as time passing, if it is not a way the world could be, then there is no contrast case with which to make sense of what it would be to misrepresent the world as being this way.5 A different approach, popular amongst passage anti-realists, is to argue that per- ception tells us that time is passing, not by representing the present as changing, but through perceptions of ordinary change and movement exhibiting what we might think of as _intrinsically dynamic phenomenology_. According to this type of account, perceiving an object to change does not just involve our perceptual systems repre- senting that object to have different properties, or be at different locations, at different moments, but also _adding on_ what Dainton calls a “distinctively dynamic sensible appearance” (2012, p.127) or, as Paul puts it, representing that object in “some suit- ably animated or flowing way” (2010, p. 352). Not realising that this dynamic quality is something produced by our own minds, we mistakenly come to believe that this feature belongs to the world, and so come to believe that we are experiencing time passing. The idea that this quality is something that the perceptual system adds to our perceptions of the world is justified by an appeal to a phenomenon known as the phi illusion (Kolers & von Grünaun 1976): subjects presented with two dots in quick succession and in slightly different locations do not have an experience of one dot followed by another, but of a single dot moving continuously through space.6 The fact that two static inputs can give rise to a perception of a single ‘animated’ object is taken as evidence that this dynamic quality is a product of the perceptual system. The final step is to say that what occurs in the phi case also occurs when we perceive real change or movement. According to the passage anti-realist the world consists of nothing but static inputs, all there is to an object moving is for it to occupy different positions at different times. However, as with the phi case, if objects are at different positions within a short enough time frame, our perceptual system presents them as having a dynamic appearance. One way of challenging this type of approach is to argue that, in fact, there is no good reason to think that the perceptual system really does add any extra property to perception of change (Hoerl 2014). While I find this line of thought persuasive, I 5 See Hoerl (2014) for more details on this type of worry. 6 Although _phi_ and related phenomena have been described many times in the philosophical and psycho- logical literature on temporal perception, I repeat these details here as they shall be important for argu- ments I make later in this section.N. Young 2624 1 3 want to argue here that even if we grant that our experiences of moving or changing objects are augmented in the manner just outlined, perceiving ordinary change cannot be the experiential ingredient that tells us time is passing. First, it is unclear how a phenomenological feature can be dynamic _in and of itself_. Even if experiences of moving or changing objects possess some feature that experi- ence of unmoving or unchanging objects lack, it is not obvious what makes this extra quality _dynamic_ in such a way that it leads to a belief in temporal passage. On a stan- dard representationalist view of perception, phenomenological features either super- vene on, or are identical to representational content (Chalmers 2004; Crane 2009). The distinctive experiential quality that we enjoy when looking at a ripe tomato or a fire engine either supervenes on or is identical to the perceptual content ‘is red’. However, it does not seem as though we can say very much more than this about the character of phenomenal properties. The quality red does not seem to be ‘dis- tinctively’ or ‘suitably’ anything other than…red. While we might have _associations_ with phenomenological redness (rage, danger. passion), the quality, in and of itself does not seem to be telling us anything more about the world than the presence of red. Similarly, while we might, for obvious reasons, associate the phenomenal character of changing objects with dynamism or animation, it is hard to see why such a quality would be dynamic, or suggestive of temporal passage, in any stronger sense.7 Third, there is the question of what to say when, as it were, the music stops. We can have perceptual experiences of nothing moving or changing at all –imagine sitting stock still, staring straight ahead, hearing only the steady hum of your central heating. Time does not seem to grind to a halt in these sorts of situations.8 I want to be careful here and distinguish the suggestion I am making from a stronger one. It is sometimes suggested (for example Torrengo 2017, p.12) that _all_ types of conscious state come with a sense of time passing. I think this is too strong, or at least requires some further argument to prove that it is so. We also often speak of situations in which we ‘lose track of time’, such as when we are thoroughly engaged in some activity like play- ing a sport or a musical instrument (e.g. what in psychology is referred to as a flow state) or when we are lost in thought.9 The challenge for accounts based on change perception is to explain why we can, at least sometimes, experience the feeling of the passing of time when there is no change around to be perceived. One response to this type of difficulty can be found in the following remarks of Prosser: While it is tempting to think that in the complete absence of sensory experience it could still seem to one that time was passing, it is not clear that one would ever be free from awareness of something—one’s own thoughts, if nothing 7 Farr makes somewhat similar objections to intrinsically dynamic qualia in his 2020. 8 O’Shaughnessy says: “Even in the situation of total lack of change in the objects of perception, change continues– within. However frozen the perceptible world may in fact be, the internal clock of conscious- ness ticks on.” (2000, p. 61). Borges’ short-story _The Secret Miracle_ (1962) provides a vivid description of what an extreme form of this type of experience might be like. 9 I will return to moments in which we lose track of time in Sects. 6 and 7.Agents of change: temporal flow and feeling oneself act 2625 1 3 else. For example I assume one would generally have an ever-changing sense of how long one had been in sensory deprivation (2016, p. 180)10 The idea here seems to be that we never experience true stillness, because even when we are not perceiving change, we continue to experience changes in our mental life. However, this seems to require two commitments which we might be reluctant to endorse. First, it seems we must say that, unlike perceptual experience, conscious thought, or some other experientially apparent aspect of our inner mental life, is _always_ changing. But this is, at least, not obviously true. Paul, for example, seems to think that we can experience unchanging thoughts: if we were in an entirely static environment where there were no contrasts between property instances (_this would have to include no contrasts with_ _respect to properties of my thoughts_), then it would seem to us as though time were standing still. And, indeed, I think this is a very plausible supposition. (Paul 2010, p. 355, emphasis mine) Intuitions might go either way here, but in the absence of an argument one way or the other, we might doubt whether Prosser’s suggestion is on firm foundations. Second, even if we grant that our inner lives are ever-changing, for this to explain why our experience of passage continues when the perceptual world stands still, introspective awareness of these changes would need to be _very_ similar to perceptual experiences of change, and this should certainly give us reason for pause. As we have seen, according to views which rely on intrinsically dynamic phenomenology, when an object is perceived as having contrasting properties in quick succession, it takes on an animated appearance, and this appearance is _a product of the perceptual system_. For changing thoughts to provide an experience of passage in the absence of per- ceived change, an inner analog of this perceptual augmentation would be required. That is, when a conscious thought has contrasting properties in quick succession the mind must add a dynamic appearance to it, just as it does when we perceive an object with contrasting properties.11 However, while there are accounts on which introspection is a form of ‘inner sense’ (e.g. Armstrong 2002, ch. 15; Nichols and Stitch 2003), even their proponents would say that there are significant differences 10 While Prosser’s view (2016, ch. 5) of why perception of change seems dynamic is somewhat differ- ent to Paul and Dainton’s, I think the issues raised in this section apply to his account as much as theirs. Hoerl (2014) thinks that all three positions can be criticised in a broadly similar way. Prosser also makes clear that he thinks that experiences of change are only part of why time seems to pass (2016, p. 179), and suggests that “the sense of temporal motion” (p. 187) and the sense that the future seems “subjectively open” (p. 198) might also contribute. However, he does not consider the sense of temporal motion to be part of our “immediate phenomenology” but is “encountered when we think about the past and present” (p. 177–178), suggesting that does not contribute to it seeming, experientially, as if time passes. And, as I make clear when considering Callendar’s (2017) view in Sect. 5, a sense of the future being open is not equivalent to an experience of passage. 11 It might be objected here that we can be aware of our thoughts changing in some other way which is not so similar to perceiving change. However, for this sort of argument to be made to work, not only would we require an account of how we experience thoughts as changing, we would also have to show why this new sort of change experience also has a dynamic phenomenology suggestive of passage.N. Young 2626 1 3 between perception and introspection, and, other than saving the idea that dynamic phenomenology leads to a belief in passage, it does not seem as if we have any reason to think dynamic phenomenology is added to aspects of our inner life. Recall also that the idea of added dynamic phenomenology in perception is motivated by the _phi_ illusion, but it is difficult to even make sense of what an inner analog to _phi_ could be. Can we make sense of the idea that two thoughts with contrasting properties could be mistaken for a single changing thought? Moreover, the idea that thoughts are experi- enced as having a dynamic appearance which they do not in fact possess is in tension with the widely held view that introspective awareness provides infallible, or close to infallible, knowledge of the aspects of our minds that we have access to (e.g. Shoe- maker 1994; Wright 2000). Even if we accept that perceptions of change possess an illusory element, it is much harder to make the analogous case for inner experience. In this section we have seen that two of the most obvious ways of cashing out the idea that perceptual experience tells us that time passes face serious difficulties. While this is not meant to be a knock down argument against the idea that a belief in time passing could come from perception, it does give us a reason to consider whether other, non-perceptual, forms of experience might do a better job. There is more to sensuous first-person experience than perceiving, and there is no obvious reason to think that the experiential ingredient which tells us time passes must be per- ceptual, or, in the case of mental change, quasi-perceptual. The discussion so far has also given us reason to consider other interpretations of what it means for experience to tell us that time is passing. There is no reason to think that a ‘direct’ representation of the present changing, nor the idea that there can be inherently dynamic phenom- enology, would fare any better if they were transposed to some other, non-perceptual form of experience. In the remainder of this paper, I will put forward a new account of how we come to believe that the present changes. The next section outlines an argument recently put forward by Hoerl on how a certain structural feature of perceptual experience can be said to _invite_ a belief in a privileged present moment. In Sects. 5, 6, and 7 I argue that _agentive experience_ invites the belief that the present moment is changing. 4 Experience as inviting belief Embedded in the belief that the present moment is changing is –obviously– a belief that there is a present moment. In this section I will outline a proposal put forward by Hoerl (2018a) as to how a certain structural feature of perceptual experience elic- its this more simple belief. Doing so will serve two purposes. First, the account I put forward in the next section can be thought to compliment to Hoerl’s: his aim is to explain how our belief in a privileged present moment is elicited by perceptual experience; my aim is to show how the belief that this moment _changes_ is elicited by agentive experience. Second, Hoerl’s argument as to how perceptual experience can be thought of as _inviting_ a belief in the privileged present will serve as a model for how I propose agentive experiences elicit a belief in the changing present. Hoerl’s starting point is an intuition that he identifies a number of writers on time and experience as having (e.g. Balashov 2005) which he summarises as follows:Agents of change: temporal flow and feeling oneself act 2627 1 3 “it appears somehow essential to the nature of perceptual experience that it is the seeming disclosure of things as they are in the present” (2018a, p. 128). As in the case of temporal passage one might be attracted to a very straightforward way of fleshing out what this seeming amounts to: perhaps perceptual experience represents, or presents, its objects as present?12 Strikingly, Hoerl goes in precisely the opposite direction, arguing that a distinguishing characteristic of perceptual experi- ence is that it does not present its objects as having _any_ sort of tensed character and that it is in fact this “lack of temporal viewpointedness” (p. 130) which is the basis for our belief that what we perceive is occurring at some metaphysically privileged moment: When we experience an event, we experience that event to the exclusion of oth- ers that happen earlier or later, but nothing in our experience signals that this is due to anything to do with us… there is a sense in which what we experience is just the bare occurrence of that event. Thus, the nature of our experience invites instead the thought that the fact that we experience this event and not the others must have a metaphysical explanation: for example, that it is just this present event that is real or in some other way metaphysically special, whereas past and future events are not (p. 146). Why would a lack of tensed content invite the belief in a privileged present? Hoerl argues for this point by comparing temporal and spatial experience. In visuo-spatial perception… part of what is manifest in the experience itself is how what a subject can see depends on where she is located: In so far as she can distinguish between the location of the objects seen and the location from which they are seen, she can also grasp the dependence of the visibility of those objects, and of the way they figure in her experience, on her own position. In this way, the character of her experience itself reveals to the subject the reason why she perceives the particular objects she does and not others, and that this is some fact about her (ibid.) What objects we see and the way those objects look, depends upon the point in space from which we are looking at them. Moreover, this dependence is _manifest_ in the experience: what we see reveals some fact about our own location in relation to those objects. Spatial experience, in and of itself, therefore provides an answer as to why _these_ objects are being perceived in _this_ way: you are seeing this ice cream, and viewing it in profile, because you are located _here_. By saying that we lack a temporal viewpoint Hoerl is saying that temporal experience lacks an equivalent feature: we do not experience whatever event we experience from another point in time (in the way we view objects from another point in space), but neither do we experience ourselves as being at _the same point in time as the event_. At any instant, we simply experience the “bare occurrence” (ibid., p.145) of some event, and temporal experience does not 12 Something like this account is put forward by Kriegel (2015a).N. Young 2628 1 3 make manifest why we are perceiving this event rather than any other. Given that the reason why we are experiencing this event rather than that one does not seem to depend on anything to do with us, we are led to believe, Hoerl argues, that it depends on some fact about the world. That is, according to Hoerl, we arrive at our belief in the privileged present abductively: at any instant of experience we are _invited to_ _believe_ that the event we are experiencing is somehow metaphysically privileged. Notice that this account of our belief in a privileged present _could_ be extended so as to account for a belief that the present moment changes. At any instant we can reflect on how perceptual experience is, was, and will be, and realise that there have been other instants which have seemed, and will be instants which will seem, as if they are metaphysically special. An obvious way of explaining why this is so would be to say that what is present changes. There is nothing wrong with this style of explanation _per se_: it shows a clear route between a particular structural feature of perceptual experience and our belief in passage. On the other hand, if we want to preserve the idea that a belief in the changing present is derived from some sort of experience of change, we might want to continue searching.13 Not to do so would be tantamount to giving up on the idea that there is a distinct _feeling_ of passage at all. The account I will now propose retains the idea that there is such a feeling, but argues that it is an agentive rather than a perceptual experience of change. 5 Experiencing oneself as a source of action. Here is an example of an event which, in contrast to those that Hoerl focuses on, you would experience as having something ‘to do with you’. Standing in the park, you move your head forward and take a lick of the ice cream you are holding. The licking event is experienced as to do with you for an obvious reason: you experience yourself as its _cause_. A number of philosophers have argued that there is something it is like to act, a _phenomenology of agency_, which accompanies our deliberate bodily movements and other types of action besides. In his 2011 Horgan et al. identifies two closely connected features of felt agency. The first of these is the “what-it’s-like of self as source”: You experience your arm, hand, and fingers as being moved by you yourself— rather than experiencing their motion either as fortuitously moving just as you want them to move, or passively experiencing them as being caused by your own mental states. You experience the bodily motion as generated by yourself (ibid., p. 79). and the second, “core optionality”: 13 Indeed, Hoerl does not intend his approach to be an explanation of our belief in temporal passage (2018b, p. 13).Agents of change: temporal flow and feeling oneself act 2629 1 3 Normally when you do something, you experience yourself as freely perform- ing the action, in the sense that it is up to you whether or not to perform it. You experience yourself not only as generating the action, and not only as generat- ing it purposefully, but also as generating it in such a manner that you could have done otherwise (ibid., p. 80). Similar ideas can be found elsewhere (e.g. Kriegel 2015b, ch. 2; Nida-Rümelin 2018; Strawson 2010),14 and I take what Horgan to be referring to here to be fairly intui- tive: we experience ourselves as the instigators of our actions, and while it might well be possible that we can experience ourselves as acting for reasons, the experience of acting does not seem to be as of introspectively bearing witness to these sorts of mental states causing us to act. We also experience these actions as something it is within our power to stop and start at any time, even if it would be irrational to do so, such as if we were being threatened to do something at gunpoint (Horgan 2011, p.81). Both aspects are obviously missing in cases of involuntary movement such as when a tap on the knee causes one’s leg to jerk. We can also note that both of these features are present _throughout_ the performance of an action. Taking a lick of your ice cream is not ballistic: you do not initiate the action and then observe your head to move forwards and your tongue to stick out. Rather, you feel yourself to be the action’s source, and feel as though you could abort your licking, at every instant that the bodily movement is occurring. In the next two sections, I will argue that our sense of ourselves as the source of our actions is an experience of change, one which invites the belief that the present moment is changing. Before that, it is worth noting that some authors have suggested that our experience of ourselves as acting freely is responsible for another common sense belief about time: the idea that the future is open. In a short section of his 2017, Callender, referring to core optionality, says: Part of what it is to be an agent is to have this sense of freedom, a sense that other future options are in some sense live… as the self “ evolves” up its world- line, it not only passes through various events and feels time to be flowing; it feels it is part of a worldwide surge of ‘ actualization’ of one path over others. The sense of passage is intimately connected with the sense that the past is settled and the future open. (p.260–261) It is true that the sense of temporal passage and the sense of an open future are con- nected. The idea that the present is changing is not far off from the idea that we can _contribute_ to how it changes, thereby affecting the future. However, it should be stressed that the two can come apart and should be dealt with separately. It is per- fectly coherent to believe that time is passing, and that experience tells us that this is so, without believing that the trajectory of unfolding events can be altered; we might believe that the present moment is changing but that _how_ it will change is inevitable. Callender himself makes clear that he considers a belief in passage and a belief in the 14 There might well be more to agentive experience than these two features. See Mylopoulos and Shepherd (2020) for an overview of other possibilities.N. Young 2630 1 3 open future to be separate components in our manifest image of time. Indeed, prior to this passage he sets out at length his own account of how the belief in passage arises, in which agentive experience does not figure at all.15,16 Finally, as we shall see in Sect. 7, the account I will put forward as to how a belief in the changing present arises rests on the idea that feeling oneself as a cause includes a certain sense of _limitation_. 6 Effecting change and feeling passage In Sect. 3 we saw that attempts to show how perceptual experiences of change lead to a belief in passage face difficulties: it is hard to make sense of how the pres- ent moment can be perceptually represented as changing (either veridically or non- veridically), and accounts which rely on our perceptions of ordinary types of change have difficulty explaining why time continues to seem to pass when no change is perceived, as well as why the phenomenological character of changing objects should lead to a belief in passage in the first place. I want to suggest now that the sense of causing one’s actions provides what perception cannot: a pervasive experience of change which is not vulnerable to questions as to how to make sense of its veridical- ity, does not rely on intrinsically dynamic phenomenology, and allows for the feeling of passage to persist, even when no object is being perceived as moving or changing. Feeling oneself as the source of an action, can be thought of as a unique type of change experience. Compare what it is like to watch the droplet of ice cream run down the ice cream cone with what it is like to move your head forward to take a lick. In the first case you are experiencing the world to change, in the second you are _experiencing yourself as changing the world_. You are not simply witnessing your head to change its location, but sensing yourself to be the instigator of the change. This seems a reasonable interpretation of what Horgan is describing when he talks about feeling your active body “as being moved by you yourself” (2011, p. 79). Simi- larly, in consciously bringing some mental image to mind you experience yourself as changing your inner world, and in shifting your attention, as changing what is at the centre and what is at the periphery of awareness. Such experiences provide a particu- larly intimate experience of change: it is not change that is experienced as belonging to the world around you but something experienced ‘from the inside’, something fundamental to what it is like to experience oneself as an agent. In a moment, I will argue that sensing oneself as effecting change makes a good candidate for the experiential ingredient which tells us the present moment is chang- ing. Before that, however, we can note that this type of change experience is not vul- nerable to the challenges faced by the perceptual accounts of the feeling of passage outlined in Sect. 3. It was mentioned there that given the causal or counterfactual constraints on perception it was difficult to see how perceiving the present moment 15 Indeed, Callender is a deflationist about the feeling of the passing of time: “What we need to explain is why creatures like us take the apparent tripartite structure of reality to be updating, not why we actually sense this” (ibid. p.230). 16 Similarly, Ismael considers our sense of ourselves as making decisions about how to act as connected to the notion that the future is open (2012) but does not consider it to contribute to a feeling of time passing.Agents of change: temporal flow and feeling oneself act 2631 1 3 as changing could be veridical, but neither did it seem as if we had the conceptual resources to make sense of it as an illusion. Our sense of ourselves as causing actions is a state of ourselves that we are aware of through introspection, which, as men- tioned, is a faculty often supposed to provide privileged access to, or perhaps even infallible knowledge of, aspects of our own minds. This allows for the difficulties suffered by the direct perceptual account to be side-stepped. Whether or not we really are causes of our own actions in the way experience presents us as being can be left as an open question, the world and all our actions within it might be fully deter- mined after all, but it is difficult to see how we could be mistaken that we _experience_ ourselves as causes, in the same way that it is difficult to imagine how we might be mistaken that we are experiencing red or feeling pain. Neither does the account I am putting forward here rely on a commitment to intrin- sically dynamic phenomenology. I am not arguing that the experience of instigating change has an extra animated element laid on top of it; rather, as we shall see in the next section, experiencing oneself as effecting change provides an explanation as to how the belief in passage arises which is similar to Hoerl’s explanation as to why we believe in a privileged present. However, we might wonder whether identifying our sense of ourselves as sources with the feeling of the passing of time is vulnerable to a version of a problem suffered by the accounts of Dainton and Paul outlined in Sect. 3. There, I argued that accounts based on change perception have a problem explaining why time does not seem to stop passing whenever we stop perceiving change. Similarly, we might ask, if the feeling of the passing of time is our sensing ourselves acting, why does time not seem to stop passing when we stop acting? Does it not seem as though time continues to pass even at those moments when we simply sit around and do nothing? One response to this challenge would be to say that the sense of ourselves as caus- ing change is an experiential constant. We feel ourselves as effecting change at every instant we are conscious. To make this type of argument, I suggest, is much easier than claiming that we always (perceptually or quasi-perceptually) witness change. While we typically think of actions in terms of pronounced bodily movements, not all movements are pronounced, and not all actions are bodily moments. First, the moments in which we are not purposefully controlling our bodies are less common than we might think. Even as you sit reading these words you are still purposefully keeping your body rigid enough for it not to slide out of your chair, and are moving your eyes in such a way as to facilitate reading, rather than letting them glaze over or drift around the page. We might think that these more subtle movements carry with them a faint, “recessive” (Bayne 2008, p. 184) sense of agency. Nida-Rümelin goes as far as saying that even actions such as dancing in an unplanned and spontaneous manner, or reacting quickly and unreflectively to a potential danger are still experi- enced as “actively brought about” by oneself (2018, p. 64). There are also _mental_ _actions_ to consider. The body could be entirely at rest and yet we can still deliberately deliberate on a problem, purposefully bring a memory to mind, or privately curse in inner speech. Consider also the experience of actively directing and maintaining one’s attention: you can shift your focus from the ice cream that you are holding to the feeling of your feet on the ground, or effortfully maintain your attention on whatN. Young 2632 1 3 your companion is saying, despite the noisy park’s myriad distractions.17 There is no reason to doubt that experiences of self as source accompany these inner actions, and it is therefore easy to explain why the feeling of the passing of time can persist even when the body is entirely still. We might think, then, that experience is _relentlessly_ _agentive_: as long as we are conscious we are always either changing where our body is in space, changing the landscape of our mind, or changing where our attention is directed. While I am sympathetic to this idea, I will not argue for it further here. This is because there is a second, more modest, response available to the challenge we are currently considering. Recall that the problem for approaches based on change per- ception was not to explain why it always seems as if time is passing but rather to explain how experience could have a dynamic flavour, even when we are not wit- nessing objects changing. We can say, therefore, that there are moments in which we are enjoy a feeling of time passing due to mental actions or shifts in attention, even when our bodies are entirely inactive, while also accepting that there are moments in which we do not experience ourselves as the source of any type of action and so do not involve a feeling of time passing. Indeed, if we think of those occasions about which we might say ‘I lost track of time’, they would seem to be cases in which we _do_ _not_ have a sense of our movements or thoughts as being brought about by ourselves. Consider moments at which one is thoroughly absorbed in some skilful activity, such as playing a musical instrument, or fully engrossed in a daydream.18 Each can be characterised not only as an occasion on which we lose track of time, but also as one in which we have little or no sense of causing our thoughts or movements: if we know the sonata well we do not have to think where to place our fingers on the piano keys, rather our movements are guided by a mixture of rote memory and immediate response to stimuli; if we are lost in a daydream we do not seem to actively change our mental landscape but instead one thought brings forth the next, which brings forth the one after that.19 If this is correct we can say that identifying our sense of instigat- ing change with the feeling of time passing, not only avoids the problems of accounts based on perceiving change but a neat answer as to why our feeling of time passing varies depending on the situation. For reasons that will become clear in a moment I want to take care in describing the character of these types of cases. In particular, we should be careful not to confuse _not experiencing oneself as effecting change_ with _experiencing oneself as not effect-_ _ing change_. Absorbed activity and daydreaming, I suggest, are cases of the former not the latter. In other words, whenever I do experience myself as an agent I experience myself as effecting change. Consider again Horgan’s description of what agentive experience is _not_ like. You do not experience parts of your body as “fortuitously mov- ing just as you want them to move, or passively experiencing them as being caused by your own mental states” (2011, p. 79). If this were how our sense of agency was 17 See Watzl (2017, ch. 11) for a discussion on shifting attention and agentive awareness. 18 See Dorsch (2015) for a discussion of the role of agency, or lack thereof, in different types of daydream- ing and mind-wandering. 19 Losing track of time is considered one of the defining characteristics of a flow state (Nakamura & Csikszentmihalyi 2014, p.90).Agents of change: temporal flow and feeling oneself act 2633 1 3 we would indeed experience ourselves as subjects but not causes. But just as this is not an apt description for what it is like to act, neither is it an apt description of what it is like to be engrossed in an action or stream of thought. Daydreaming or playing the piano is not a case of ‘stepping back’ and observing a series of automatic thoughts or movements. Rather, when one is truly absorbed, the sense of agency evaporates. 7 Having to act and believing in passage Experiencing oneself as an agent is to experience a particularly intimate form of change, which pervades all, or at least a large proportion of, the instants at which we are conscious. Here, I shall argue that, while not being a representation of a changing present, this type of experience invites us to believe that there is such a thing in much the same way that Hoerl says that a lack of temporal view pointedness invites a belief in the privileged present. Section 4 showed that there is a contrast between spatial experience of objects and temporal experience of events. Our spatial experience makes manifest both why we are perceiving the objects we are perceiving and the way that they are perceived: I am seeing _this_ ice cream in profile because I am viewing it from _this_ spatial viewpoint. Temporal experience is not viewpointed in the same way and so does not make mani- fest why we are perceiving this event at this particular instant. In agentive experience we find analogs to both. In sensing myself as a generating my actions, the reason why my body is moving the way it is is made experientially manifest: I sense myself as the source of my licking, and as not taking up any of the other options that are avail- able to me (dropping my ice cream, tightening my grip so as to crush it etc.). But what agentive experience does not make manifest, is _why I invariably find myself as_ _effecting change_. It is this negative feature of agentive experience, I suggest, which provides a basis for the belief in a changing present. As mentioned above, even if there are moments in which we do not experience ourselves as agents, such as in a daydream, whenever we _do_ experience ourselves as agents, we always experience ourselves as doing _something_. That is, we never experience ourselves as agents who are not effecting change.20 Whatever the reason for our being in such a state, it is not revealed in experience: I do not sense myself as the cause of my having to act. If I did, it would seem to me as if I had the option to be entirely inactive: I would not only feel as if I could stop licking my ice cream at any instant, but also that, at any instant, I could put myself into a state in which I am _not_ effecting change. As any insomniac will attest, we cannot choose to shut ourselves down in the way that we can a laptop, and neither do we experience ourselves as being able to do so.21 20 This point is somewhat similar to Sartre’s idea that we are condemned to be free (1948/2007): we always find ourselves as having to make some choice. That is, to effect some sort of change. 21 Of course it is possible to act in such a way as to _facilitate_ an inactive state: we close our eyes in the hope that sleep will come, but this is not to choose to stop acting so much as to choose to act in a way which might stop you acting at some later point.N. Young 2634 1 3 The fact that inactivity never seems to be an option shows that we do not experi- ence ourselves as the source of our being active. Neither does experience present our active state as being caused by something other than ourselves. Consider how some involuntary movements have what we might call _felt causes_. If you squirm when I poke you in the ribs you do not experience yourself as the source of your squirming, but neither is it a mystery to you as to why your body is moving like that: you experi- ence the poking to be the cause of your squirming. Being in an active state, however, does not seem to have a felt cause. We do not experience ourselves as being prodded into action but are instead presented with the bare occurrence of our effecting change in a way similar to how Hoerl describes perception as presenting us with the bare occurrence of events.22 As we saw, Hoerl argues that this negative feature of temporal perception invites the belief that there is a privileged present moment. I want to sug- gest that, similarly, not experiencing the cause of our being active invites the belief that this present is changing. We saw that on Hoerl’s account, by not providing an answer as to why we are experiencing this event rather than any other, perception invites the idea that there must be some metaphysical fact independent of ourselves which explains this. Simi- larly, agentive experience does not present us with a reason as to why we are active, and so can be thought of as inviting the idea that there must instead be some fact independent of ourselves which accounts for this instead: if I am not causing it, and do not experience anything else as causing it, it must be something else. If this line of thinking is correct then the next step is to show that we believe this something else to be a changing present. Here, my explanation is slightly different to Hoerl’s. While his relies on an implicit comparison between what we perceive and what we remember or anticipate, mine relies on accounting for a feature which is ever present in agentive experience: the pervasive, intimate, and immanent experience of change that is feeling ourselves act. Every instant at which we catch ourselves doing, or thinking, or attending, is an expe- rience of change, and while we have an explanation as to how this change unfolds -we sense ourselves as guiding it- we do not have an explanation as to _why there is_ _change at all_. Whenever we perceive change we can ask why it is occurring: Why is the ice cream melting? And if the change is not experienced as something to do with us, we come to believe that it is due to some fact about the world. If I am not caus- ing the ice cream to melt, something (or things) other than me must be responsible: some property of the ice cream itself, some feature of the environment, etc. Given that agentive change is there at every instant we experience ourselves as acting, this leads us to believe that it does not have a local, contingent cause, but a global one. 22 While the question as to the actual cause of our being active is interesting, it is not something I need to address here. Any answer would most likely appeal to the fact that our brains are perpetually active sys- tems. The patterns of neural activation which underlie our perceiving, thinking about, and responding to the world do not cease for as long as we are awake, and so it is unsurprising that we always find ourselves doing something. If we were aware of this neural churning as causing our activity, then, according to the argument I am making here, we would not come to believe that the present moment is changing. However, my proposal does not rely on experience withholding the _correct answer_, only that it provides _no answer_. Even if agentive experience were misleading and presented a cause for our being in an active state which was not the real cause, this would be sufficient, I suggest, to prevent it inviting belief in a changing present.Agents of change: temporal flow and feeling oneself act 2635 1 3 That is, our having to act seems to require a metaphysical explanation: the change we experience must be a feature of reality itself. According to Hoerl, perceiving the bare occurrence of some event or other leads to the belief that the instant of percep- tion is somehow special; on my account it is the bare occurrence of our own agentive restlessness which leads us to believe that that instant is in some special sort of flux. Experiencing the irresistible flow of our actions leads us to believe that we are being swept along by a changing present. 8 Conclusions I have argued here that if we are looking for an aspect of experience which elicits the belief in the passage of time, then our sense of ourselves as sources of action is a promising candidate. The feeling of time passing is the feeling of acting, of effecting change. Unlike direct perceptual accounts, this approach is not vulnerable to worries regarding its veridicality or lack thereof, and, unlike accounts based on perceptions of ordinary change, provides a uncomplicated account of why the feeling of time pass- ing persists even when we are not witnessing change occurring in the world around us. While the experience of generating change is not a representation of the present moment changing, there is a clear route from this type of experience to a belief in a changing present moment. The fact that whenever we experience ourselves, we experience ourselves as effecting change, coupled with the fact agentive experience does not make manifest to us _why_ this is the case, provides an explanation as to how agentive experience invites a belief in a changing present. Acknowledgements This research was funded by the Department of Philosophy “Piero Martinetti” of the University of Milan under the Project “Departments of Excellence 2018–2022” awarded by the Ministry of Education, University and Research (MIUR) and the project “Time and Emotion” (ID-35533) of the University of Milan. Support was also received from the project “Rethinking and communicating time (CHRONOS)”, PID2019-108762GB-I00 of the Spanish Ministry of Science and Innovation. For feedback on previous versions of this material, thanks to all participants at the ‘The First International Confer- ence on Time’ at Osaka City University in November 2019, and Giovanni Merlo’s Time Reading group held online in November 2020. Thanks also to two anonymous referees for their extremely helpful com- ments, and my colleagues at the Centre for Philosophy of Time at the University of Milan, in particular Davide Bordini, Dave Ingram, Cristian Mariani, and Luca Marchetti. Special thanks to Giuliano Torrengo for his many detailed comments on many aspects of this paper, as well as Natalia Panchieri and Bryan O’Donovan, both of whom have helped me think more clearly about the issues that I write about here. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/ licenses/by/4.0/.N. Young 2636 1 3 References Armstrong, D. M. (2002). _A materialist theory of the mind_. Routledge Arstila, V. (2018). Temporal experiences without the specious present. _Australasian Journal of Philoso-_ _phy_, 96(2), 287–302 Balashov, Y. (2005). Times of Our Lives: Negotiating the Presence of Experience. _American Philosophical_ _Quarterly_, 42(4), 295–309 Bayne, T. (2008). The phenomenology of agency. _Philosophy Compass_ , 3(1), 182–202 Borges, J. L. (1962). _The secret miracle. Labyrinths: Selected stories and Other Writings_, 88–94 Callender, C. (2017). _What makes time special?_. Oxford University Press Cameron, R. P. (2015). _The moving spotlight: An essay on time and ontology_. Oxford University Press Oxford Chalmers, D.(2004). The representational character of experience. In: A. Bardon (Ed.), _The Future of the_ _Philosophy of Time_ (pp. 153–181). New York: Clarendon Press Correia, F., & Rosenkranz, S. (2018). _Nothing to come: A defence of the growing block theory of time_ (Vol.395). Springer Crane, T. (2009). Intentionalism In: McLaughlin,B., Beckermann, A., & Walter, S. (Eds.). (2009). _The_ _Oxford handbook of philosophy of mind_. Oxford University Press Dainton, B.(2012). Time and Temporal Experience. In: A. Bardon (Ed.), _The Future of the Philosophy of_ _Time_. New York: Routledge Davies, P. (1995). _About time: Einstein’s unfinished revolution._ Penguin Deng, N. (2013). Fine’s McTaggart, temporal passage, and the A versus B-debate. _Ratio_, 26(1), 19–34 Dorsch, F. (2015). Focused daydreaming and mind-wandering. _Review of Philosophy and Psychology_, 6(4), 791–813 Farr, M. (2020). Explaining temporal qualia. _European Journal for Philosophy of Science_, 10(1), 1–24 Hoerl, C. (2014). Do we (seem to) perceive passage?. _Philosophical Explorations_,17(2),188–202 Hoerl, C. (2018a). Experience and time: transparency and presence. _Ergo, an Open Access Journal of_ _Philosophy_, 5 Hoerl, C. (2018b). Transparency, Time, and Perspective: Reply to Richardson, Carter and Lee. https:// philosophyofbrains.com/wp-content/uploads/2018/09/Brains-blog-reply_CH.pdf Horgan, T. (2011). The phenomenology of agency and freedom: Lessons from introspection and lessons from its limits. _Humana Mente_, 15, 77–97 Ingram, D. (2018). _Thisness Presentism: An Essay on Time, Truth, and Ontology_. Routledge Hohwy, J., Paton, B., & Palmer, C. (2015). Distrusting the present. _Phenomenology and the Cognitive Sci-_ _ences_, 15(3), 315–335. https://doi.org/10.1007/s11097-015-9439-6 Ismael, J.(2012). Decision and the open future In: A. Bardon (Ed.), _The Future of the Philosophy of Time_. New York: Routledge Kolers, P. A., & von Grünau, M. (1976). Shape and Colour in Apparent Motion. _Vision Research 16/4_, (pp. 329–35) Kriegel, U. (2015a). Experiencing the present. _Analysis_, 75(3), 407–413 Kriegel, U. (2015b). _The varieties of consciousness_. Oxford University Press Latham, A. J., Miller, K., & Norton, J. (2020). Do the folk represent time as essentially dynamical?. _Inquiry_ , 1–32 Latham, A. J., Miller, K., & Norton, J. (2021). Is our naïve theory of time dynamical? _Synthese_, 198(5), 4251–4271 Lee, G. (2014). Extensionalism, atomism, and continuity. In: Oaklander, L.N. (Ed.), _Debates in the Meta-_ _physics of Time_, (pp. 149–173) Leininger, L. (2013). On Mellor and the future direction of time. _Analysis_, 74(1), 1–9 Miller, K. (2019). Does it really seem to us as though time passes?. In: _The Illusions of Time_ (pp. 17–33). Palgrave Macmillan, Cham Miller, K., Holcombe, A., & Latham, A. J. (2020). Temporal phenomenology: phenomenological illusion versus cognitive error. _Synthese_, 197(2), 751–771 Mozersky, M. J. (2015). _Time, Language, and Ontology: The World from the B-Theoretic Perspective_. Oxford, UK: Oxford University Press Mylopoulos, M., & Shepherd, J.(2020). The experience of agency. In: U. Kriegel (Ed.), _The Oxford Hand-_ _book of the Philosophy of Consciousness_ (pp. 164–168). Oxford University PressAgents of change: temporal flow and feeling oneself act 2637 1 3 Nakamura, J., & Csikszentmihalyi, M. (2014). The concept of flow. In: _Flow and the foundations of posi-_ _tive psychology_ (pp. 239–263). Springer, Dordrecht Nichols, S., & Stich, S. P. (2003). _Mindreading: an integrated account of pretence, self-awareness, and_ _understanding other minds_. Clarendon Press/Oxford University Press Nida-Rümelin, M. (2018). Freedom and the Phenomenology of Agency. _Erkenntnis_, 83(1), 61–87 O’Shaughnessy, B. (2000). _Consciousness and the World_. Oxford: Oxford University Press Paul, L. A. (2010). Temporal experience. _The Journal of Philosophy_, 107 (7), 333–359 Phillips, I. (2013). October). XII—Perceiving the passing of time. In: _Proceedings of the Aristotelian_ _Society_ (Vol. 113, No.3_pt_3, pp. 225–252). Oxford, UK:Oxford University Press Phillips, I. (2014). The temporal structure of experience. In: _Subjective time: The philosophy, psychology,_ _and neuroscience of temporality_, 139–158 Prosser, S. (2013). Passage and perception. _Noûs_, 47(1), 69–84 Prosser, S. (2016). _Experiencing time_. Oxford University Press Sartre, J. P. (1948/2007). _Existentialism is a humanism_. (C. Macomber, Trans.). Yale University Press Shardlow, J., Lee, R., Hoerl, C., McCormack, T., Burns, P., & Fernandes, A. S. (2021). Exploring people’s beliefs about the experience of time. _Synthese_, 198(11), 10709–10731 Shoemaker, S. (1994). Self-knowledge and “inner sense.” _Philosophy and Phenomenological Research_, 54, 249–314 Skow, B. (2015). _Objective becoming_. Oxford University Press,USA Strawson, G. (2010). _Freedom and Belief_ (Revised Edition). Oxford: Oxford University Press Torrengo, G. (2017). Feeling the passing of time. _Journal of Philosophy_, 114(4), 165–188 Watzl, S. (2017). _Structuring mind: The nature of attention and how it shapes consciousness_. Oxford University Press Wright, C.(2000). Self-knowledge: the Wittgensteinian legacy. In: C. Wright, B. C. Smith, & C. Macdon- ald (Eds.), _Knowing Our Own Minds_. Oxford: Oxford University Press Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. </writing_example_2> <writing_example_3> Beauty in Use:  Agentive Phenomenology and the Aesthetics of Design  Introduction Good design seems like something that we can aesthetically appreciate. There are 'design classics' like the Bialetti coffee pot, the 1994 Swiss Railway Clock, and the iPod, as well as very many examples of bad design –TV remotes cluttered with meaningless buttons, phone chargers with plugs so heavy they slip from the socket. While these examples highlight the visual aspects of design, we argue that the aesthetic experience derived from interacting with designed objects should not be overlooked. Current philosophical accounts of design aesthetics, despite their focus on function, pay insufficient attention to experiences of use. Relying on recent work in the aesthetics of dance and the aesthetics of games, we will argue that appreciating design is a type of playing or dancing with a technical artifact. Yet, we will also rely on agentive phenomenology to point out that the aesthetics of design differs from the aesthetic of dance and the aesthetics of games in some crucial respects.  In Section 1 we introduce Parsons and Carlson’s theory of Functional Beauty as a means of explaining how function factors into our perceptual appreciation of good design. On their account, a designed object which looks as if it functions well (looks “fit for function”) then it will, in virtue of this, have aesthetic merit. However, we will argue that a complete aesthetics of design cannot be entirely perception based. A laptop keyboard might look functionally beautiful, but not actually function very well once users begin to type. In Section 2, we will examine Forsey's Kant-inspired view according to which aesthetic appreciation of design is fundamentally a matter of making a judgement about how perfect it is as regards its function. Both perception and use can inform these judgements, meaning that this approach fares better at explaining how invisible design might be appreciated.  Yet, we argue that even this account does not fully capture the aesthetic experiences arising from the use of designed objects. Use need not just be a means of assessment; there are also cases when use itself is the object of aesthetic attention.  While the aim of this first half of the paper is to show that user experience has a part to play in the aesthetics of design, the aim of the second is to provide an account of what an aesthetically valuable user experience is. In Section 3, we introduce what we call 'the question of difficulty'. On the one hand, we use designed objects either to enable certain actions or make them easier: one cannot drive nails without a hammer, a dishwasher makes washing dishes easier and more efficient. On the other, easy actions are not automatically aesthetically satisfying: flicking a light switch or opening a drawer are not typically aesthetically pleasing interactions. Nguyen's agentive aesthetics of games provides one answer to the question of difficulty: aesthetically pleasing gameplay arises when the game's challenges perfectly match and push the player's skills to their limit - a "harmony of capacity". Some similarities to aesthetically pleasing interactions with designed objects may suggest that the aesthetics of design is a matter of playing with technical artifacts. Yet, aesthetically satisfying use of technical artifacts rarely involves actions that push the user's skills to their limit in the way that a difficult game might. In light of this, we suggest a more promising way of explaining the relation between difficulty and the aesthetics of use can be found by adapting Montero's account of aesthetic effortlessness in dance. Just as the aesthetic experience of the dancer, according to Montero, is rooted in the sense of performing a difficult task with ease, we argue that the user experience relevant to the aesthetics of design is similarly grounded. In this sense, the aesthetics of design is a matter of dancing with technical artefacts. In Section 4, we explore the agentive phenomenology and intentional structure of user experience to differentiate the aesthetics of design from both the experience of play and the experience of dance. We argue that designed objects support actions integrated into our daily lives and connected to broader goals and purposes. Unlike games and dance, which involve framed actions within a "magic circle" (Huizinga 1955), design allows us to appreciate actions that are part of our everyday activities. By recognising these differences, we highlight the specific aesthetic qualities that well-designed objects contribute to our interactions with the world. In sum, our aim is to highlight a dimension of design aesthetics that current philosophical theories have either not focussed on at all (in the case of Parsons and Carlson) or not focussed on to the extent to which it merits (in the case of Forsey). However, we are not arguing that use is the only way to appreciate design, suggesting that existing theories should be rejected because they do not in themselves account for it. Rather, we consider the ideas we introduce in what follows as a supplement to existing theories, but an important and hitherto largely overlooked one. Let us finish this introduction with three preliminary considerations about terminology and scope.  First, the notion of aesthetic appreciation that we deploy draws on Gorodeisky’s (2023, p. 85) idea that “aesthetic value is to be understood in light of a capacity for a feeling of pleasure, a feeling through which we grasp aesthetic value, are responsive to what it merits, and exercise our agency”. From this perspective, aesthetic appreciation amounts to a rewarding experience in which one feels a peculiar pleasure that the experienced object deserves.  Second, the notion of beauty that is already at work in our title matches what Zangwill calls “verdictive beauty”, that is, pure aesthetic merit regardless of the specific ways in which it is realized in the objects that have it. Zangwill aptly distinguishes this from what he dubs “substantial beauty” which is instead a specific way of realizing aesthetic merit, as opposed to, say, the sublime.  When Parsons and Carlson (2008, xii-xiii) introduce the notion of “functional beauty”, they have in mind the same “verdictive” meaning of ‘beauty’ that we are adopting, as this passage clearly shows: “this one expression, ‘Beauty’ is used, as philosophers occasionally still use it, to mean ‘aesthetic appeal in general’, rather than the specific sort of aesthetic appearance it typically suggests”. Third, as we are developing an aesthetic account of design based on use, our focus is on objects that can be used, in the sense that they need to be acted upon by their user to perform their function. What we mean here is roughly the same as Lowe's distinction between utensils and machines (2014). Utensils are objects that require our active engagement to fulfil their function, and include things like knives, bicycles, and keys: if the user takes their hands off a knife it will cease cutting, if they stop pedalling they will slow to a halt. Machines, on the other hand, can perform their function without need for constant contact with the user. An oven keeps heating once it has been turned on, an electric fan will keep spinning until it is switched off. While Lowe's distinction tracks our everyday understanding of utensils and machines for the most part, it does not always do so. We would not normally think of a bookshelf to be a machine, but it will keep your books suspended above the ground without user involvement, and so would come out on Lowe's account as a machine. There are also hybrid cases. The iPod, for example, is utensil-like in that users have to move their finger or thumb around the click wheel to select a track, but machine-like in that once the track is selected music will play without further input. Finally, it seems right to extend the notion of utensil to include objects which we make perceptual and cognitive use of like signs, maps, books and websites. Although we do not manipulate these objects with our physical bodies, we nevertheless use them to gain information about the world. Our arguments concerning the aesthetics of use will apply to these things as well as utensils like knives or bicycles, and hybrid cases like iPods. 1. The Aesthetics of Invisible Design 2. 1 Functional Beauty Perceptual experience, visual experience in particular, plays a very prominent role in philosophical accounts of design. Aesthetic appreciation of objects like the Bialetti and Swiss Railway Clock depends, at least in part, on their visual appearance. However, we will argue in this section that accounts solely based on vision, or even perception more generally, are not the end of the story when it comes to appreciation of design: we should also consider what it is like to use an object. Using an object is of course closely related to the function of the object, and function does play a role in existing accounts of design appreciation. For example, on Parsons and Carlson's notion of functional beauty, "knowledge of [...] function can play a role in our perception of and hence our aesthetic appreciation of that object"(2008 p. 90). Visually appreciating a Bialetti would therefore differ from visually appreciating a sculpture because knowledge of the coffeepot's practical purpose would feed into one's visual experience, in a manner not possible in the case of sculpture, given its lack of practical purpose. Parsons and Carlson propose three ways function translates into visual aesthetic qualities (2008, pp. 97-100). First, an object can look fit for function, appearing well-suited for its purpose by having expected features plus specific elements that enhance functionality. This quality emerges in objects like muscle cars, which look fast through their standard features and additional elements like spoilers. Second, an object can display elegance and simplicity through appearing to have no extraneous or irrelevant features, consisting primarily, or entirely of elements that contribute to its function. A sleek stainless-steel stove exemplifies this, with its standard features like glass door and burners creating an elegant appearance. Third, an object can look as though it is capable of performing its function while at the same time possessing features that appear to contradict this capability, creating a surprising, vibrant look. The Beverly Hills lawn chair creates this sort of "visual tension" (ibid. p. 99) with its three-quarters S-shaped frame that appears too insubstantial to support a person's weight. Functional Beauty places the aesthetics of design firmly in the perceptual realm. Parsons and Carlson write mainly about visual perception, but we can apply similar principles to other senses. Roberts extends this approach to tactile perception, which we'll examine below. Yet a solely perceptual view of design aesthetics may be incomplete. Some aspects of design appreciation are not captured by perception alone. 1. 2 Design Is Not Just What It Looks Like We can begin to see this by noting that not all talk of good and bad design is based around the visual appearance of objects. It is easy to think of examples of noticeably bad design: capsule coffee machines which make it difficult to remove used capsules, mobile phones made from slippery materials are easy to fumble or drop. Moreover, it is easy to think of cases in which how visually pleasingly functional something looks comes apart from how pleasing it is to use. Apple's butterfly keyboard is an eloquent example of this. In March 2015 Apple released an iteration of its very successful MacBook laptop line. In the accompanying press release Apple emphasised the thinness of the keyboard, describing it as "34 percent thinner"  and "dramatically thinner and highly responsive" than previous models (2015). While the butterfly keyboard did make the laptop thinner, a common criticism after its release was that its shallowness compromised the typing experience. These feelings are exemplified by comments on the popular website Laptopmag.com, describing it as "the keys were so shallow that it felt like typing on a hard surface" and finding it "uncomfortable for extended typing sessions due to its low key travel and lack of tactile feedback" (2019). Despite the laptops' sleek and minimal design, these usability complaints, along with reliability concerns, led to the butterfly keyboard being phased out of Apple’s lineup by May 2020. This failure brings out what a perception-based account like Functional Beauty is lacking, as Steve Jobs himself said in 2003: “Most people make the mistake of thinking design is what it looks like. People think it’s this veneer — that the designers are handed this box and told, ‘Make it look good!’ That’s not what we think design is. It’s not just what it looks like and feels like. Design is how it works.” (Walker, 2003) 1.3 Tactile Functional Beauty How can we appreciate good design if it is beyond vision? One way is to extend functional beauty, or any perception based account, to perceptual modalities other than vision. If good invisible design cannot be seen, perhaps it can be felt, or heard, or smelled? Roberts has argued that the functional beauty account can be extended to include haptic touch. He argues that just as we can see an object as functionally beautiful, we can feel it that way too. He proposes that "an object feels fit for its function when it is perceived through touch as having no contra-standard properties for an object of its type, and as having variable properties, all of which are indicative of a high degree of functionality" (p. 56). Roberts illustrates this with the example of a perfect snowball:  a faultless sphere that fits neatly in the palm, with just enough heft and texture for throwing without risk of injury to its victim. Not being unusually large or small, and lacking any embellishment of form (ibid.) These tactile qualities indicate the snowball's fitness for its function, leading to what Roberts calls a "pleasing haptic quality" of "feeling fit for function" (ibid.). He extends this concept to other objects, such as bicycles, where haptic exploration reveals qualities that contribute to functional beauty but remain visually inaccessible (p. 57). For example, touch allows appreciation of a bike frame's lightness or smooth gear operation, indicating its fitness for efficient riding. Haptic touch also permits detection of objects unfit for their function. While users could not visually perceive the butterfly keyboard's design flaws, they could feel them through use. But even if we accept this haptic extension of functional beauty, it doesn't seem to fully capture our experience of design. While Roberts' account helps us understand how we might appreciate the tactile qualities that indicate fitness for function, it still focuses on perceptual properties of the designed object rather than the experience of use itself. Think about using a well-designed smartphone app. The pleasure of using it does not derive from how the screen feels under our fingertips. Or consider driving a finely-tuned sports car. The pleasure of driving it goes beyond the tactile sensations of gripping the steering wheel or pressing the pedals. In both the app case and in the car cases, there's an aesthetic quality to the interaction itself. Finally, there are designed objects which we use without touching them, such as the Swiss Railway Clock or Harry Beck’s London Underground Map. All of this suggests that to fully account for the aesthetics of design, we might need to look beyond perception altogether and consider the aesthetics of use itself. 2. The Aesthetics of Use 2.1 A Kantian Approach Whereas Parsons and Carlson's focus on perception, Forsey (2013) develops a Kantian approach to design aesthetics based around assessments of how well objects function. She draws on the distinction between free and dependent beauty, which Kant's (§16,65) draws as follows: “The first presupposes no concept of what the object ought to be; the second does presuppose such a concept and the perfection of the object in accordance therewith”. While free beauty is judged solely on form, dependent beauty is judged by considering also what Forsey (p. 163) characterizes as “the purpose which determines what the thing is to be" . She applies this notion of dependent beauty to design, arguing that the aesthetic appreciation of designed objects necessarily involves assessing how well they fulfil their functions, considering both their purpose and how effectively they realise that purpose. Forsey illustrates her application of dependent beauty to design through her analysis of coffee pots. The aesthetic judgement of a coffee pot involves assessing both its form (such as its shape and material) and how these elements contribute to its function (pp. 181-182). For instance, an octagonal shape might be judged beautiful not just for its visual appeal, but for how it allows easy opening even with wet hands. This exemplifies what Forsey terms "teleological style" - the specific way an object realises its purpose (p. 167). Such judgments go beyond mere visual appreciation, incorporating how well the object works in practice. Forsey argues that many qualities of a designed object can only be fully assessed through use. She illustrates this point in her discussion of judging design excellence: This judgement of perfection cannot be made by looking at the coffee-pots alone. Designed objects are meant to work, but to ascertain whether they do, and whether they are good things of their kind, requires us to use them or put them to the test: is the coffee any good? Can we lift the lid to serve it? Does the pot leak or the metal corrode? Some of these virtues and deficiencies will not be apparent to the eye alone, and this means that judgements of design excellence...require more than observation...(p. 183) This passage shows that Forsey’s approach can accommodate non-perceptual features of design quite straightforwardly. Such features will show up when the object is put to work. We can see this through Forsey’s comparison between her own coffee-pot and “his”: But his coffee-pot, I want to claim, has flaws that are hidden behind that newness and shine, that detract from its beauty. First, brass conducts heat, and each time you reach for the handle, or put your finger on the lid, you burn yourself. Bakelite remains cool. Second, the sleek rounded design makes it very hard to unscrew the two halves, especially if you already have soapy hands. My octagonal pot turns as easily as a nut in a wrench, whether wet or dry. Third, the conical shape of his means the opening of the top pot is too narrow to fit even a small hand in to clean it, whereas mine, as wide at the top as at the bottom, welcomes a quick scrub. (p. 181 – p. 182) While the aluminium of Forsey’s coffee-pot has started to corrode (it looks less fit for function that it used to), and its finish is no longer shiny (its formal properties are not what they were), using it allows for various hidden qualities to show up: coffee can be made without burning one’s fingers, it opens more easily, and it is more easily cleaned. 2.2 The Pleasure of Use Unlike Parsons and Carlson’s Functional Beauty, Forsey’s account makes room for use. However, on her account using an object is merely a means of assessing how well an object performs its function. We suggest that this approach does not fully capture the role that using an object plays in design appreciation. We can begin to see this by noting that although Forsey is right to say that certain virtues of her coffee-pot do not show up if we just look at the thing, they can be revealed by seeing someone else use it. Watching whether someone else singes their fingers while making coffee will be just as informative as singeing one’s own fingers. However, we suggest, simply observing an object being used leaves something important out: the experience of interacting with that object.  Forsey's account, in this respect, bears some resemblance to Nguyen’s notion of “harmony of solution”, which he describes as the aesthetic quality arising from how well a solution fits a given problem or challenge. He writes: "In a beautiful Chess move, there is a harmony between the move and the situation it addresses. Here is a trap; there is an elegant resolution. Call this the harmony of solution. The harmony of solution is strictly a harmony between the solution and the obstacle; it makes no explicit reference to the actor or their capacities. The harmony of the solution is available to both spectator and player." (p. 102) This concept could apply to design in terms of how well a designed object solves a problem. It might explain our appreciation of clever or elegant design solutions. For instance, we might admire the elegant simplicity of a paperclip or the ingenious design of a Swiss Army knife.  The harmony of solution, however, has limitations when it comes to explaining the aesthetics of interaction between user and object. It doesn't focus on the experience of using the object. As Nguyen points out, this harmony is available to both users and observers, making it not specific to interaction. While it may contribute to our overall appreciation of a design, it does not capture the specific aesthetic experience of interaction we're seeking to explain. Consider the following example. For the last few months a professional violist’s favourite viola has been in the repair shop. Finally, it is returned to her and she begins to play. As she does so, she appreciates the unique tone that only her viola can produce, this is something she has missed in the months she has had to play her backup viola. Moreover, she has never been able to play the back up quite as well: its poorly cut bridge does not allow for the nuanced playing made possible by her favourite, and its rougher fretboard meant she can never move her fingers as smoothly as she can with her favourite. These features show up for the violist as she is playing, but both would also be available to an audience. Discerning listeners would judge her favourite to be superior to her back up. However, there is an aspect of use which is available only to violist herself: the experience of playing her viola. It is not simply that she can play better with her favourite, it is that her favourite feels better to play. Using an object allows for aesthetically appreciable interactive experiences, mere observation does not. Similarly, negative experiences of use can be hidden from observers: one needs to use the butterfly keyboard personally to discover how unsatisfying it is to type on.  It is in fact common to value objects (utensils) in terms of what it is like to use them. Consider again Roberts’ example of the perfect snowball. While he focuses on how the snowball's perceptual properties - its spherical form, size, and texture - indicate fitness for throwing, this analysis overlooks an aesthetic dimension. The snowball's aesthetic value emerges not just from how it looks and feels in the hand, but from the experience of actually throwing it. A snowball that seems perfectly formed can still provide an unsatisfying throwing experience, just as one that appears imperfect might soar perfectly through the air. The aesthetic appreciation of objects like snowballs involves not just their static properties, but the shape of the actions they elicit. Consider also the sorts of product review which litter Youtube and the internet more generally. A review of a car will make reference to how it looks, efficacy, ease of use etc. but they will also often tell you how it feels to drive. Does it feel satisfying to turn the steering wheel, accelerate etc.? These considerations suggest the following: the aesthetics of design, at least when it comes to utensils like a viola or a fork, has an agentive component. Use is not merely a means of assessing how well an object performs its function, but an aspect of aesthetic appreciation in its own right. In the following section we will show how an aesthetics of interaction with designed objects should be understood.  3. The Aesthetics of Interaction 3.1 Appreciating actions Even if one accepts that there are experiences of interaction, and that they can be more or less pleasurable, we can still ask whether this pleasure is aesthetic. Why think that the viola player’s appreciation of how it feels to play her favourite is more similar to commonly agreed upon aesthetic pleasures like taking in a work of art or looking out over a landscape, rather than garden variety pleasures like taking a bath or tasting sugar? GIven Gorodeisky’s (2023)account of aesthetic value, why does a viola or a bicycle deserve the pleasure we take in riding it whereas sugar does not deserve the pleasure we take in eating it? While we have no knock down argument to this challenge, the idea that one’s own actions can be objects of aesthetic appreciation is not uncommon in contemporary aesthetics. First, Montero writes about dancers aesthetically appreciating their own movements:    “it is natural, at least for dancers, to talk of experiencing beauty proprioceptively. A dancer, during a rehearsal onstage—a situation in which there are no mirrors—may claim that a certain movement or position is beautiful or […] complain that the beauty, or whatever other aesthetic quality he or she is aiming at producing, is lacking: 'The movement is too abrupt’; 'The line is ugly'; 'I'm not feeling the connections' are all phrases that roll naturally off a dancer's tongue" (2006 p. 232). Montero argues that proprioception can function as an aesthetic sense, allowing dancers to appreciate the beauty or ugliness of their own movements without needing visual feedback. She defends this idea against objections related to the private nature of proprioceptive experience and its supposed inability to support aesthetic disagreements. Montero also extends her argument to suggest that observers can make aesthetic judgments through kinesthetic empathy, indicating that proprioception's role in aesthetics extends beyond self-awareness. Nguyen goes one step further, arguing that one can aesthetically appreciate one’s own actions, and that this can occur both in everyday life but especially in games: Games, I’m suggesting, are process-centered. When I make aesthetic judgments in response to a game, I am ascribing aesthetic properties and qualities to my own processes—to the actions and activities that I perform in response to the game. (Nguyen 2020,  p. 142) Phenomenologically, we can differentiate actions from mere movements in terms of our sense of agency: the sense that we are the source of our movements. There is a phenomenological difference between your leg moving because a doctor has tapped it with a hammer, and you deliberately moving your leg. What is lacking in the former but present in the latter case is a sense of you yourself as causing the movement. We can see from the passage below that Nguyen also believes that we can aesthetically appreciate our own mental actions as well as bodily ones:  Fixing a broken car engine, figuring out a math proof, managing a corporation, even getting into a bar fight—each can have its own particular interest and beauty. These include the satisfaction of finding an elegant solution to an administrative problem, of dodging perfectly around an unexpected obstacle. These experiences are wonderful—but in the wild, they are far too rare. Games can concentrate those experiences. When we design games, we can sculpt the shape of the activity to make beautiful action more likely. And games can intensify and refine those aesthetic qualities, just as a painting can intensify and refine the aesthetic qualities we find in the natural sights and sounds of the world. (2020, p. 13) As with bodily movements we can differentiate mental actions, such as deliberately working through a mathematical proof, from mere occurrent thoughts, such as when a thought suddenly ‘pops’ into your head unbidden.  The analysis of aesthetic appreciation of actions has direct relevance for understanding design. Design shapes the possibilities for action, with well-crafted objects enabling particular types of engagement with the world. Physical artifacts like violas and coffee-pots can elicit aesthetically pleasing actions from users through their form and construction—the way a viola's well-cut bridge enables nuanced playing, or how an octagonal coffee pot allows for secure grip even with wet hands. The design of utensils also can sculpt the shape of the activity to make beautiful action more likely reinforcing how objects not only facilitate function but structure the quality of the actions they make possible. This capacity, shaping both function and the quality of action, is not limited to physical artifacts but extends to cognitive artifacts - objects whose primary function involves conveying information through being perceived, such as pages of text or webpages.  The aesthetic quality of engagement manifests here too: text laid out with appropriate spacing and typeface selection enables fluid, pleasing reading, while poorly designed layouts create disjointed, aesthetically unsatisfying experiences. In both physical and perceptual cases, design shapes the aesthetic potential of the actions that the designed objects enable.  3.2 The Question of Difficulty While the agentive aesthetics of design and the agentive aesthetics of games are similar in many ways, we will argue here that they differ in at least one respect, which can be brought out by considering the relationship between agency and difficulty.  Questions about aesthetic pleasure as it relates to difficulty arise for both games and interactions with designed objects. In games, we might ask why people enjoy difficult games (Juul 2013; Kretchmar 2012). With designed objects, the question is slightly different: what exactly is the relation between aesthetically pleasing use and level of difficulty. On the one hand, designed objects are created to either make an otherwise impossible task possible, or to make a possible task easier: telescopes and microscopes function to make otherwise invisible objects visible; washing machines function to make washing clothes more convenient. On the other, our lives are filled with mundane, easy actions that we do not typically find especially aesthetically satisfying.As we shall see, Nguyen’s solution for the question of difficulty in games does not provide a satisfactory answer to the question of difficulty in design. Nguyen uses the concept of harmony of capacity to explain why people enjoy difficult games. He describes this as the aesthetic quality arising from how well one's overall capacities match the demands of a task, especially when pushed to their limits: "This is an experience, not just of a particular action's fitting the requirements at hand. It is an experience of your whole self fitting the task. It is the experience of your abilities, worked at their maximum, just barely making it." (p. 109) Consider the exhilaration of playing Tetris at a high level. As the speed increases, every action you make will be right at the limit of your capacities, you are managing to place each block just in time. As Nguyen points out, this sort of harmony is rare in everyday life and is normally only encountered in games.  He explains that many everyday tasks are either too easy, too difficult, or too repetitive to allow for an experience of this type of harmony. Moreover, it's difficult to quantify or standardise the "edge of ability" for everyday objects. While the harmony of capacity might be applicable to some specialised design contexts, it doesn't fully capture the range of aesthetically pleasing interactions we're interested in for everyday design objects. This highlights our first point of difference between game aesthetics and design aesthetics: the relationship between difficulty and aesthetic satisfaction in using designed objects cannot be straightforwardly explained by Nguyen's notion of the harmony of capacity. 3.3 Aesthetic Effortlessness in Design Interactions We suggest that the best way of accommodating the relationship between aesthetically pleasing experiences of interaction and ease of use is through Montero's ideas of aesthetic effortlessness. In her work, she aims to characterise what it is for something to appear effortless, and explain why effortlessness is aesthetically appealing. Montero does not equate something looking effortless with something looking easy, as can be seen in the following passage: "When we attribute effortlessness to bodily movements in dance, it seems that at least in many cases we also see the work as, in some sense, difficult; we see it as difficult, yet appearing easy." This apparent paradox of difficult actions appearing easy extends beyond dance. In music, for instance, Bach's fugues can seem, in a sense, obvious to a skilled pianist, despite their complexity and cleverness. These examples highlight that aesthetic effortlessness often involves a perceived ease in the execution of objectively challenging tasks. To explain why we find this quality aesthetically appealing, Montero adapts Herbert Spencer's (1852) idea of 'superfluity of fitness': "If the movements we dub as effortless are movements that would be for us difficult to perform yet appear to be performed with reduced effort, then part of the reason why effortless action is attractive could be that it reveals a superfluity of fitness." (Montero, 2012, p. 140) The concept of 'superfluity of fitness' illuminates why effortlessness is aesthetically appealing. When we perceive an action as effortless, we implicitly recognise that the actor possesses capabilities that exceed the demands of the task at hand. This surplus of ability allows for the execution of even complex actions with apparent ease, creating a harmonious relationship between the actor and the action. The aesthetic appeal of this effortlessness stems from its suggestion of mastery and control. When we observe a violinist effortlessly playing a challenging piece or an athlete gracefully performing a difficult manoeuvre, we experience a sense of admiration. We feel a pleasure that these actions seem to fully deserve. This response arises not merely from the successful completion of the task, but from the perceived harmony between the actor's capabilities and the action's requirements. While Montero's account focuses on the perception of effortlessness in others, our analysis requires a shift to the first-person experience of effortlessness. This leads us to the concept of agentive effortlessness, which includes proprioceptive effortlessness as one of its forms. Agentive effortlessness refers to an agent's experience of their own actions as effortless, even when performing objectively difficult tasks. In the context of design interactions, the concept of agentive effortlessness can explain how users derive aesthetic satisfaction from using well-designed objects. This type of effortlessness does not concern the ease of simple actions or mundane interactions discussed earlier, but rather the specific experience of performing tasks in a manner that reduces perceived effort while still requiring a degree of engagement. The application of agentive effortlessness to design interactions reveals two distinct types of experiences: 1. Expert-oriented agentive effortlessness: tools like a well-balanced chef's knife. Such objects allow users to develop and express high levels of skill. The knife's design facilitates smooth, precise cutting, enabling experienced users to perform complex tasks with apparent ease. The aesthetic satisfaction derives from the user's sense of their own expertise, amplified by the tool's design. This aligns closely with Montero's original concept, where the superfluity of fitness is most apparent in skilled performances. 2. Novice-oriented agentive effortlessness: devices like the original iPod. Here, the design allows even novice users to perform what would otherwise be complex tasks (managing a large music library) with ease. The aesthetic pleasure stems from the user's sense of accomplishing something sophisticated effortlessly, despite their lack of expertise. This extends Montero's concept by suggesting that good design can create a sense of superfluity of fitness even in users who lack high levels of skill. In both cases, the aesthetic quality of the interaction stems from the user's experience of effortlessness, which is facilitated by the object's design. This effortlessness reflects a perceived abundance of ability relative to the demands of using the object, echoing Montero's concept of superfluity of fitness. This account of agentive effortlessness in design interactions offers several advantages over Nguyen's concept of harmony of capacity: 1. It explains aesthetic satisfaction across a broader range of interactions, from simple to complex. While Nguyen's harmony of capacity focuses on experiences at the edge of one's abilities, agentive effortlessness can account for aesthetic experiences in both challenging and everyday interactions. 2. It focuses on the user's subjective experience of ease, rather than the objective match between task difficulty and user ability. This allows for aesthetic experiences in interactions with well-designed objects that don't push users to their limits. 3. It directly addresses how object design contributes to the aesthetic experience. The concept of agentive effortlessness explicitly considers how the design of an object can facilitate a sense of effortlessness in the user, whether expert or novice. 4. It aligns with the design goal of creating objects that feel effortless to use, while still allowing for the development and expression of skill. This dual consideration of both ease and skill development is particularly relevant for design contexts. In sum, by adapting Montero's concept of aesthetic effortlessness to the realm of agentive experience in design interactions, we can account for the aesthetic satisfaction derived from interacting with well-designed objects. This framework accommodates both the simple pleasure of effortless everyday interactions and the more complex satisfaction of skilful use, providing an explanation of the relationship between difficulty, ease, and aesthetic pleasure in design that aligns with and extends Nguyen's aesthetics of agency. 4. Agentive Phenomenology and Intentional Structure 4.1 Nguyen's Harmony of Action and Agentive Phenomenology In Section 3, we explored how Nguyen's concept of harmony of capacity, while illuminating for game aesthetics, has limited application to the aesthetics of design. Nguyen also introduces the notion of "harmony of action," which he describes as the aesthetic quality arising from how well an individual's actions fit the demands of a situation. Unlike the harmony of capacity, Nguyen believes that harmony of action can be experienced both in games and in real life: "When you time a jump just so in Super Mario Brothers; or when you figure out, during a rock climb, that you need to slide your hips over just enough to balance on that tiny nubbin of rock, you're experiencing more than the harmony of solution. You're experiencing your agency and action as fitting the demands of the environment." (Nguyen 2020, p. 108) Earlier, he provides a real-world example: "A car swerves into my lane and I react smoothly, dodging perfectly. Afterward, I am struck by the gracefulness of my own instinctive response." (ibid., pp. 101–102) Harmony of action can be applied straightforwardly to design aesthetics. Design choices directly affect our ability to achieve that harmony. Take the example of avoiding an obstacle in a car: good design choices about steering sensitivity and suspension response might let us experience our defensive driving as perfectly matched to the situation. Poor design choices about these same elements might make identical manoeuvres feel uncontrolled and awkward. This suggests that one way design shapes aesthetic experience is by enabling or preventing such harmony. What we have said in the previous section as for aesthetic effortlessness in design interaction might be traced back to the broader notion of harmony of action. However, there is a difference here that merits attention. While it is reasonable to think that something like the harmony of action factors into both real-life and in-game interactions, the way in which actions fit the environment differs in each type of case. This difference is reflected in our agentive phenomenology—that is, in the way we experience ourselves as agents acting in the world. Agentive phenomenology refers to the subjective experiences associated with exercising agency, particularly during intentional actions (Mylopoulos and Shepherd 2020). When we act intentionally, we typically experience ourselves as initiating our actions, feeling in control, and able to start or stop our actions at will, even if it might be unwise to do so. This sense of agency is absent in cases of involuntary movement, such as when a tap on the knee causes our leg to jerk reflexively. In both games and real life, we can have this sense of agency, but the nature of our agentive experience can differ depending on the context. In games like Super Mario Brothers, our actions are shaped by the game's design. The environment is controlled, and the challenges are predefined. Timing a jump perfectly to overcome an obstacle fits precisely within the game's immediate demands and offers instant feedback.The player's actions are confined within the "magic circle" of the game (Huizinga 1955), a boundary that separates the game world from ordinary life. Within this context, the reasons for our actions and the way they fit the environment are provided by the game's rules and objectives. In contrast, real-life actions occur in more complex environments embedded in a web of considerations. Our decisions and actions are influenced by personal values, social norms, legal obligations, and long-term goals. The way our actions fit the environment in real life is therefore more intricate, and this complexity is reflected in our agentive phenomenology, which includes various aspects such as the feeling of initiating an action, the sense of control over one's movements, the felt possibility of being able to do otherwise, and the awareness of acting for reasons (Mylopoulos and Shepherd 2020). These experiences can vary depending on the context in which we act, leading to differences in how we experience our own agency in games versus real life, as we shall now see. 4.2 Chromatic Illumination, Purposiveness, and Means-Ends Teleology To understand how our agentive phenomenology differs between real life and games, we can consider two specific aspects: purposiveness and means-ends teleology. Purposiveness refers to the experience of acting on purpose, of engaging in actions that are goal-directed. When we act intentionally, we often have multiple purposes, both broad and specific. For example, when we enter our office in the morning, our general purpose might be to start the workday, but this involves specific purposes like turning on the computer, checking emails, and preparing for meetings. These purposes influence our experience of acting, even if we are not explicitly thinking about them at every moment. Means-ends teleology involves understanding our actions as means to achieve certain ends. It includes recognising how our desires and beliefs combine to provide practical reasons for action. For instance, we might desire to finish a project and believe that working uninterrupted for an hour is necessary to achieve that goal. This belief-desire combination shapes our experience of acting as we organise our activities to meet our objectives. These aspects of our agentive experience are often present in the background of awareness, influencing our actions without being the focus of our attention. In a dialogue with Nida-Rümelin, Horgan introduces the concept of chromatic illumination to describe how these background purposes and meanings colour our agentive experience without being explicitly represented in consciousness: "Often, many of the purposes informing one's actions affect the character of one's agentive experience by being implicitly appreciated consciously without actually being represented in consciousness. These purposes are phenomenologically operative as reasons for one's actions, colouring the specific character of one's agentive experience without being overtly 'before the mind.'" (Horgan & Nida-Rümelin 2020, p. 265) Through chromatic illumination, we are aware of our purposiveness and means-ends teleology as background aspects of our experience. In real-life actions, our purposes and means-ends relationships are numerous and layered, involving personal values, social norms, and long-term aspirations. This complexity contributes to the way we experience ourselves as agents, as our actions are connected to various facets of our lives. In games, the purposiveness and means-ends teleology of our actions are different. The purposes and means-ends relationships are typically defined by the game's rules and objectives, confined within the game's magic circle. While we still act on purpose and use actions as means to achieve ends, these purposes are more straightforward and less embedded in our broader life context. We are aware of these differences because the background purposes in games are less interconnected with personal values and long-term goals. Therefore, the way in which actions fit the environment in games versus real life leads to differences in our agentive phenomenology, particularly regarding purposiveness and means-ends teleology. In real life, these aspects contribute to a richer experience of agency. Our actions are connected to our personal narratives and social contexts, influencing how we perceive and engage with the world. In games, the experience of agency may be more focused and confined, reflecting the structured nature of the game's environment and objectives. 4.3 The Agentive Aesthetics of Design  The differences in agentive phenomenology between real-world and game actions influence the aesthetics of design, particularly in terms of the harmony of action they afford. One main way in which the agentive aesthetics of using designed objects differs from that of gameplay lies in the richer phenomenology of purposiveness and means-ends teleology present in real-life actions, contributing to a more integrated harmony of action. Designed objects enable a harmony of action that is integral to our daily lives and connected to our purposes. When we interact with a well-designed object, such as a finely made tool or a user-friendly piece of technology, our agentive experience involves purposiveness and means-ends relationships that align with our broader goals. This alignment can make our agentive experience aesthetically pleasing, as the object helps us to act in ways that matter to us, thereby deserving the pleasure we take in it. In contrast, games, while providing enjoyable experiences, offer a different kind of harmony of action that is confined within the "magic circle," in which actions are framed and their meanings are limited to the game context. The agentive aesthetics of gameplay is shaped by the game's predefined purposes and means-ends relationships, which, although present, may not align as deeply with our everyday purposes. Therefore, while both games and designed objects involve purposiveness and means-ends teleology, the harmony of action is more integrated in the case of designed objects, leading to agentive experiences that differ from those in games. By recognising how designed objects facilitate a harmony of action in our interactions with the world, we can see how they enhance our everyday experiences. Our agentive experiences with designed objects are aesthetically pleasing because they help us act harmoniously in ways that we value in our daily lives. The agentive aesthetics of design can thus contribute to the “everyday aesthetics” (Saito 2017) in a way the agentive aesthetics of games could not. Design, from this perspective, is to games as architecture is to sculpture. Both sculpture and architecture shape volumes but the former does so in a framed way, in the “magic circle” of the art gallery for instance, whereas the latter does so in the wild. Likewise, both games and design shape actions but the former do so in their magic circles whereas the latter does so in the wild.   Evnine's (2016) hylomorphic account of actions can further illuminate this point. According to Evnine, actions can be seen as artifactual events, where the intentional form represents the agent's purposes and goals, and the material aspect involves the physical movements performed. In sum, actions are constituted by an intentional form impressed onto a matter that consists of bodily movements. In the context of design, this alignment of intentional form and material movement contributes to a harmony of action, as the designed object facilitates actions that are integrated with our purposes and means-ends relationships. For example, when using a well-designed tool, the action is not just a physical movement but is harmoniously informed by our intentions and purposes. This hylomorphic account of actions  ultimately enables us to properly differentiate the aesthetics of design from both the aesthetics of dance and the aesthetics of games. Aesthetically appreciating design is a sort of playing or dancing with a technical artifact. The aesthetics of design differs from the aesthetics of dance since the former may also consist of mental actions while the latter is based on bodily movements. Designed objects are closer to games than dance in this respect since the aesthetic appreciation of them both can be centred on actions whose hylomorphic matter consists of changes in mental content rather than bodily movements. However, the aesthetics of design differs from the aesthetics of games in virtue of the depth and complexity of its chromatically illuminated form. Games and dance are similar in this respect since they both involve framed actions. Design, instead, leads us to aesthetically appreciate actions intertwined with our life.  Conclusion This paper has explored the role of user experience in the aesthetics of design, arguing that current philosophical accounts have not fully addressed the aesthetic dimensions arising from the use of designed objects. Beginning with Parsons and Carlson's theory of Functional Beauty, we observed that while perception—particularly visual perception—plays a significant role in appreciating design, this approach is limited when considering objects where good design is not immediately apparent through perception alone. We highlighted that certain designs are such that their aesthetic value emerges primarily through interaction and use rather than perceptual features. We then examined Forsey's Kantian perspective, which incorporates function into aesthetic judgement by assessing how well an object fulfils its intended purpose. Although this view allows both perception and use to inform aesthetic appreciation, we argued that it still does not fully capture the aesthetic experiences derived from the act of using an object. Specifically, we argued that use should not be regarded merely as a means of evaluating functional efficiency but should be recognized as an object of aesthetic attention in its own right. To address this gap, we introduced the concept of agentive aesthetics, drawing on Montero's notion of aesthetic effortlessness. We proposed that the aesthetic satisfaction derived from using designed objects often stems from the experience of performing tasks that are inherently challenging yet are executed with apparent ease. This sense of effortlessness reflects a harmony between the user's abilities and the demands of the task, facilitated by the design of the object itself. Unlike Nguyen's agentive aesthetics of games, which emphasise actions that push users to their limits, we suggested that aesthetically pleasing interactions with designed objects do not typically involve actions that test the user's maximum capacities. Furthermore, we discussed how the richer phenomenology of purposiveness and means-ends teleology in real-life actions contributes to the agentive aesthetics of design. Interactions with designed objects are embedded within broader personal goals and contexts, unlike actions in games that are confined within a structured environment. This integration enhances the aesthetic experience, as the use of well-designed objects aligns with and facilitates the user's intentions and purposes in everyday life. By incorporating the aesthetic appreciation of use into the discussion on design aesthetics, we have aimed to provide a more comprehensive account that complements existing theories. Recognising the aesthetic value that emerges from the interaction between users and objects enriches our understanding of what constitutes good design. This perspective acknowledges that the agentive pleasure derived from seamless interactions is a significant dimension of aesthetic appreciation, thereby offering a more complete picture of the aesthetics of design without dismissing the importance of perceptual qualities or functional assessments. The dualism of form and function is surely the crux of the aesthetics of design and has traditionally been addressed as if the form were related to the experience of the technical artefact while the function were related to a more cognitive or pragmatic stance towards that artefact. This paper has shown that both form and function involve an experiential dimension and the dualism of form and function might be better cast as the intertwining of perceptual phenomenology and agentive phenomenology within one aesthetic experience.  References Apple. (2015, March 9). Apple Unveils All-New MacBook. Apple Newsroom (Sénégal). Retrieved October 14, 2024, from [https://www.apple.com/sn/newsroom/2015/03/09Apple-Unveils-All-New-MacBook/](https://www.apple.com/sn/newsroom/2015/03/09Apple-Unveils-All-New-MacBook/) Barranco, M. C. (2025). Design and Aesthetic Appreciation: Form, Functionality, Performativity. Journal of Comparative Literature and Aesthetics, 48(1), 23-35. Evnine, S. J. (2016). Making objects and events: A hylomorphic theory of artifacts, actions, and organisms. Oxford University Press. Forsey, J. (2013). The aesthetics of design Oxford University Press. Gorodeisky, K. (2023). Aesthetic value: The view from here. The Journal of Aesthetics and Art Criticism, 81(1), 85-86. Horgan, T., & Nida-Rümelin, M. (2020). On the satisfaction conditions of agentive phenomenology: A dialogue. In The Routledge handbook of phenomenology of agency (pp. 264-299). Routledge. Huizinga, J. (1955). Homo ludens: A study of the play-element in culture. Beacon Press. Juul, J. (2013). The art of failure: An essay on the pain of playing video games. MIT Press. Kant, I. (1972). Critique of judgement (J. H. Bernard, Trans.). Hafner. (Original work published 1790) Kretchmar, S. (2012). Competition, redemption, and hope. Journal of the Philosophy of Sport, 39(1), 101-116. Kriegel, U. (2015). The varieties of consciousness. Oxford University Press. Lowe, E. J. (2014). How real are artefacts and artefact kinds? In Artefact kinds: Ontology and the human-made world (pp. 17-26). Montero, B. (2006). Proprioception as an aesthetic sense. The Journal of Aesthetics and Art Criticism, 64(2), 231-242. Montero, B. (2016). Aesthetic effortlessness. In S. Irvin (Ed.), Body aesthetics (pp. 180-191). Oxford University Press. Mylopoulos, M., & Shepherd, J. (2020). Agentive phenomenology. In U. Kriegel (Ed.), The Oxford handbook of the philosophy of consciousness. Oxford University Press. Nguyen, C. T. (2020). Games: Agency as Art. Oxford University Press. Parsons, G., & Carlson, A. (2008). Functional beauty. OUP Oxford. Roberts, D. (2019). Haptic Functional Beauty. British Journal of Aesthetics, 59(1), 53-68. Saito, Y. (2017). Aesthetics of the familiar: everyday life and world-making. Oxford University Press. Spencer, H. (1892). Gracefulness. In Essays: Scientific, political, and speculative (Vol. 2, pp. 381-386). D. Appleton and Company. (Original work published 1852) Tracy, P., Smith, S. L., Tabari, R., & Murray, M. (2019, December 29). Apple keyboard face-off: Are the MacBook Pro's "scissor" keys really better? LaptopMag. Retrieved October 14, 2024, from https://www.laptopmag.com/features/apple-scissor-vs-butterfly-keyboard-macbook-pro von Bonsdorff, P. (2025). Design and tacit aesthetics: Design as the art of objects-in-use. Journal of Comparative Literature and Aesthetics, 48(1), 13-22. Walker, R. (2003, November 30). The guts of a new machine. The New York Times. Retrieved from https://www.nytimes.com/2003/11/30/magazine/the-guts-of-a-new-machine.html Zangwill, N. (2019). Beauty and its kinds. The Journal of Value Inquiry, 53(3), 435-451. </writing_example_3> <new_piece_to_create> A 4000 word philosophical article based on this TEXT. it must be analytic and argue for a strong and philosophically interesting conclusion.: TEXT: # Transformer Token Mechanics 1. Token relationships within the architecture Every input symbol—sub‑word unit, byte‑pair fragment, or other token—is transformed into a fixed‑length vector by an embedding matrix that is usually tied to the output projection. A sine–cosine positional encoding or a learned alternative such as rotary or relative embeddings adds location information so that order can be recovered after the permutation‑invariant attention step. Inside each transformer layer the query vector qiq_i for token ii forms dot products with the key vectors kjk_j of all tokens jj. After scaling by dk\sqrt{d_k} and applying softmax, the resulting weights αij=softmax(qi⋅kj/dk)\alpha_{ij} = \mathrm{softmax}(q_i \cdot k_j / \sqrt{d_k}) indicate the share of attention token ii gives to token jj. The weighted sum ∑jαijvj\sum_j \alpha_{ij} v_j of the corresponding value vectors vjv_j becomes token ii’s message from that head. Because attention is computed in parallel heads, different heads can focus on orthogonal subspaces—one may learn syntactic distance while another specialises in long‑range coreference. A residual connection adds the head output back to the input, and layer normalisation stabilises the statistics before a position‑wise feed‑forward network applies a ReLU\mathrm{ReLU} or GELU\mathrm{GELU} transformation. Repeating this pattern across dozens of layers builds representations that capture morphemes, phrases, discourse features, and factual links while the objective remains simply to predict the next symbol. 2. Prompt interaction with the architecture Inference applies a causal mask that zeros out any score between a position and a future key, enforcing left‑to‑right information flow. All prompt tokens therefore lie in the receptive field of every prediction that follows. The continuous nature of the attention weights means that a prompt steers computation probabilistically rather than through hard branching. Prompt length interacts with the fixed context window. When the sum of prompt and generation approaches the limit, the model must truncate or compress earlier tokens, sometimes degrading performance because early information can no longer be attended to. Researchers exploit the prompt’s softness through methods such as prefix‑tuning, soft prompts, and virtual tokens that learn to imitate hundreds of real words while occupying only a few positions. Because each layer recomputes attention, the original prompt redistributes its influence at every depth, allowing even a brief instruction to persist across the entire stack. 3. Mechanics of next‑token prediction After the final attention and feed‑forward blocks, the hidden state at the last position passes through a linear map Wo\mathbf{W_o} (often the transpose of the embedding matrix, enforcing weight tying) and a softmax to yield a categorical distribution over the vocabulary. Training minimises cross‑entropy loss between this distribution and the ground‑truth next token. Gradients flow backwards through attention heads, feed‑forward layers, residual paths, and positional encodings, updating parameters so that future sequences are modelled more accurately. During generation the network is called repeatedly with an expanding context. Sampling is not limited to greedy or nucleus choice; beam search, contrastive decoding, and decode‑transform‑refine schemes alter exploration–exploitation trade‑offs. Temperature rescaling flattens or sharpens the logits, and repetition penalties or frequency penalties mitigate loops. Because attention complexity is O(n2)\mathcal{O}(n^2), long contexts present computational challenges; sparse and linearised variants approximate the attention matrix when efficiency dominates. 4. “Reasoning tokens” and problem‑solving Asking the model to emit chain‑of‑thought exposes internal computation. The visible tokens then enter subsequent attention computations, letting the model condition later steps on its own provisional conclusions. A self‑consistency technique samples multiple chains and averages the answers; reflection variants feed the chains back for critique and repair, reducing hallucination. Retrieval‑augmented generation treats citation strings as reasoning tokens that carry external facts into the context. Because these strings are indistinguishable from ordinary text, they do not change the architecture; they merely create additional nodes for attention to traverse. 5. Why next‑token prediction enables complex behaviours Minimising negative log‑likelihood across diverse text imposes information‑theoretic pressure to discover every exploitable regularity: morphology to compress character sequences, syntax to compress word order, world knowledge to anticipate plausibility, and pragmatic cues to model discourse. Multi‑head attention builds a latent space in which semantically related contexts cluster, reducing the entropy of the output softmax. Scaling studies show a power‑law relationship between loss and parameter count, data size, and compute, suggesting that deeper and wider networks steadily accumulate compressed representations of increasing scope. Capabilities that resemble reasoning or planning appear because the training corpus embeds traces of such cognition—proof steps, cause–effect explanations, and instructions. The network learns to reproduce these patterns whenever they lower expected loss. Even tasks never encountered during optimisation, such as zero‑shot translation into an obscure dialect, can surface if their solution pattern overlaps with structures already compressed in the weights. --- Takeaway Transformer token mechanics rely on universal building blocks—embeddings, multi‑head self‑attention, residual feed‑forward layers, and a causal mask—arranged to minimise next‑token loss. This simple objective incentivises the emergence of sophisticated representations, while prompts and optional reasoning strings steer the distribution without altering the architecture. As parameter counts and context windows grow, the same mechanics continue to extend the model’s behavioural range, turning statistical compression into the appearance of understanding. </new_piece_to_create> </inputs> <developer_note> In the past, you've undershot similarity. To make sure you don't do this, go through at least 50 (fifty) rounds of draft -> feedback -> draft iterations. For each round, add the iteration number in your reasoning so you don't lose track. Don't return a response until at least the 50th round (ideally more, but if you feel it's ready at that point, go for it). </developer_note> -