Experimental Aesthetics and Conceptual Engineering

Experimental Philosophy (X-Phi) is now a fully-fledged methodological project with applications in almost all areas of analytic philosophy, including, as of recently, aesthetics. Another methodological project which has been attracting attention in the last few years is conceptual engineering (CE). Its areas of implementation are now diverse, but as was the case initially with experimental philosophy, aesthetics has unfortunately been left out (or perhaps aestheticians have failed to pay attention to CE) until now. In this paper, I argue that if conceptual engineers are interested in expanding their project to the field of aesthetics, which would greatly benefit the field, then they should rely on the existing experimental work of aestheticians. Experimental philosophers have only recently started to join forces with conceptual engineers in various fields, as well as to explore the methodological implications of such an alliance. This paper goes a step further by not only arguing that CE has potential in aesthetics, but that the way to realize this potential is to piggyback, so to speak, on the work of experimental aestheticians. In other words, instead of building a CE project in aesthetics from the ground up, this paper describes the support that CE can and should derive from current experimental aesthetics, thereby making the former’s development more efficiently realizable. Furthermore, I argue that doing so would also be beneficial to experimental aesthetics. Currently, the integration of X-Phi to the wider field of aesthetics is losing ground because certain objections— notably, the objection that X-Phi cannot be of relevance to normative questions— have not been properly refuted. By pairing up with a normative programme like CE, though, experimental aestheticians should finally be able to put these objections to rest.


Introduction
Experimental Philosophy (X-Phi) is a project within analytic philosophy with substantive metaphilosophical commitments. Broadly speaking, it is committed to informing philosophical investigations with various experimental methods. The movement is now almost two decades old and has branched off in several directions since its inception, which has made it fairly difficult for metaphilosophers to agree on the specific characteristics of the movement. 1 One thing that can be said is that X-Phi has substantive metaphilosophical commitments because, among other things, it seeks to answer a metaphilosophical question: "how should we do philosophy?". As such, it can be further defined as a methodological project within metaphilosophy. Other metaphilosophical questions might include "what is the nature of philosophy?" or "what is the purpose of philosophy?". These questions can be treated individually, or not. In fact, with closer scrutiny, we would likely find that those questions are more intertwined than we realize and thus that any given metaphilosophical project is never isolated. For instance, the purpose of philosophy might depend on its nature, and its methodology on its purpose. It might even be the case that metaphilosophical progress (which ultimately is philosophical progress) can only be achieved if these various projects are developed in tandem.
An example of this in aesthetics is Meskin and Liao's (2019) paper, 'Experimental Philosophical Aesthetics as Public Philosophy'. I understand public philosophy to be another project with substantive metaphilosophical commitments. 2 It is a practice that questions the nature of philosophy-is it merely an academic practice reserved for experts with a given set of technical skills?-its purpose-is the goal of philosophy to accumulate knowledge for the sake of it or can it be socially, culturally, politically significant?-and even its methods-given the target audience, are the traditional methods of analytic philosophy, its language and its format, the optimal tools to communicate ideas?
As such, Meskin and Liao's public experiments on coffee appreciation presented in their paper are, as they argue, an instance where two projects with substantive metaphilosophical commitments were combined and benefited from one another. In their public experiment, Meskin and Liao had some of their participants describe the taste of different coffees, selected by a renowned local coffee shop, while other participants had to match the produced descriptions to the original coffees. The public event they put on produced interesting empirical information about the difficulties of communicating taste, which contributes to a larger discussion in X-Phi about aesthetic communication, but it also generated an opportunity for the participants themselves to philosophise about aesthetic issues from an experimental standpoint.
1 See Buckwalter and Sytsma (2016) for a fairly thorough volume on X-Phi, both its methodology and its areas of application. For a specific overview of experimental aesthetics, see Cova et al. (2015) or Torregrossa (2020). 2 Whether or not those who practice public philosophy see themselves as having metaphilosophical commitments is another matter.
In the same spirit, this paper looks at two specific projects with substantive metaphilosophical commitments in aesthetics: one which already has a few devoted supporters, i.e. X-Phi, and another which has yet to be taken up in aesthetics, 3 but with many proponents in other areas of analytic philosophy, namely conceptual engineering (CE). Very roughly for now, CE does not investigate what folk concepts look like, but rather investigates what our concepts should be like.
The primary goal of this paper, then, is to demonstrate that, in so far as aesthetics might benefit from CE, CE can and should draw and, thus, benefit from experimental aesthetics. My predictions, all applied to aesthetics, are as follows: (1) X-Phi and CE are compatible and (2) X-Phi can and should help CE get off the ground more effectively. Secondarily, I will also suggest that (3) X-Phi can also benefit from CE.

What is Conceptual Engineering and is it Compatible with X-Phi?
Before assessing the viability of combining X-Phi and CE in aesthetics, we must first find out whether the two separate projects are compatible. This might be an obvious truth for many in the field, and in this section I address, and agree with, their claims of compatibility. However, I do wish to reiterate some of these arguments here, at the risk of being redundant, for the sake of aestheticians who might not be familiar with these arguments. I believe this to be necessary because, as was the case for X-Phi, the uptake of CE in aesthetics has not followed the pace of other philosophical fields. As such, to evaluate the compatibility of the two projects in aesthetics, I must look, briefly, at the fields where CE has been implemented and speculate about the role it can play in aesthetics. Luckily, conceptual engineers tend to discuss the nature and the merits of CE from a generalised rather than localised point of view, so it should not be too difficult to apply their arguments to my area of choice.
As I have mentioned, X-Phi is mainly a methodological project because it offers new investigative methods for analytic philosophy. To a certain extent, so is CE. Just like X-Phi, I think that CE can be construed as an alternative to the prevailing methodology in analytic philosophy, namely conceptual analysis. 45 Briefly, CE looks at 3 At the time of writing, I could find only two instances in which conceptual engineers have engaged with aesthetics: (1) Kevin Scharp gave a talk on 'Conceptual Engineering and Aesthetic Concepts' at the Scottish Aesthetics Forum in December 2019, and (2) The British Society of Aesthetics sponsored a Connections Conference on the topic of 'Conceptual Engineering and Aesthetics' in September 2021. However, the results from these talks have yet to be published. 4 On PhilPapers, CE, X-Phi and Conceptual Analysis are all categorized under 'Philosophical Methods', which is a subcategory of 'Metaphilosophy'. 5 I understand that conceptual analysis can sometimes be construed narrowly (and historically) as the specification of necessary and sufficient conditions for the correct application of particular mental items (e.g. concepts) or linguistic items (e.g. terms). I suspect that many analytic philosophers nowadays see themselves as doing much more than describing such conditions for such items. However, I also believe that they fail to see that what they actually do has much more in common with this original sense of conceptual analysis than with the metaphilosophical projects discussed here, i.e. X-Phi and CE. What I am referring to when I discuss conceptual analysis nowadays is a broad category of philosophical methods and practices which are directly inherited from something like the original sense of conceptual analysis. I take it that any kind of descriptive, definitional or conceptual work, be it case-based, commitment-based, concepts from a normative standpoint. Instead of analysing a given concept to find out what it is, conceptual engineers ask what the concept ought to be. 6 They argue that analytic philosophy can identify deficient concepts and revise them for the better. 7 They also present many different reasons why a concept might be deficient, some practical (for instance a concept might promote social injustice by undermining certain moral or political values) and some theoretical (a concept might be inconsistent or it might even be superimposed, which is when a single concept is covering too much ground and ought to be split into distinct concepts). 8 To improve deficient concepts, conceptual engineers also suggest different strategies. For instance, we could simply abandon the deficient concept. Another strategy is to replace it altogether with another, maybe existing, concept. Or we can try to 'ameliorate' it, which can take different forms as long as the identified deficiencies are fixed. A couple of canonical examples of CE projects are Haslanger's expansive work on race and gender (2000,2012)-hers is what she would call an 'ameliorative project' (2000)-and Scharp's (2013) work on replacing the concept of truth, which he argues is inconsistent. Roughly speaking, these are some of the most central characteristics that any CE project has.
However, just like X-Phi, CE has taken many forms (and names) since its inception and pinning it down is beyond the scope of this paper (and perhaps altogether pointless). 9 Also beyond the scope of this paper is explaining what conceptual engineers mean by 'concepts'. Just like experimental philosophers discussed and tested intuitions without all agreeing on what the term referred to, conceptual engineers seem to be working with different notions of 'concept'. Machery (2017) takes concepts to be 'psychological entities' (he also refers to CE as 'conceptual explication' in reference to Carnapian explication). 10 Cappelen (2018) refers to them as 'representational' or 'expressive devices'. As such, I will not take a stand here on what conception of concepts we should accept in order to make the claim that X-Phi and CE are compatible. I hope to make a compatibility claim broad enough that it can more or less be applied to any case of X-Phi and CE. If they are any methodological 10 See Nado (2021) for a discussion of Carnapian explication and experimental philosophy.
Footnote 5 (continued) intuition-based, and so on, will constitute this category of conceptual analysis. Moreover, according to the 2020 PhilPapers survey by Bourget and Chalmers,70.92% of 1733 respondents (who were predominately analytic philosophers) accept or lean towards conceptual analysis being the most useful/important philosophical method, which is the highest percentage for any of the philosophical methods listed on the survey. 6 To make this normative aspect more salient, some philosophers also refer to CE as conceptual ethics (Burgess & Plunkett, 2013). 7 The implementation of the revised concepts post-engineering is also taken to be a part of the overall project of CE (see Cappelen et al., 2020), although no precise strategy has yet been established for that end. 8 Here, I am following the 'Theoretical Framework' put together online by the members of the Con-ceptLab, including Cappelen, Linnebo and Serck-Hanssen. Currently, the ConceptLab is the foremost research centre on conceptual engineering and as such, I am following their lead concerning the methods and aims of the project. See their website: https:// conce ptual engin eering. info/ theor etical-frame work/. Another taxonomical resource for conceptual deficiencies is Cappelen (2018). 9 Again, Cappelen (2018), especially chapter 2 provides an overview of the movement. divergences in any given instance of either project, adjustments might need to be made, but this can be done on a case-by-case basis.
At first glance, X-Phi and CE might not be an obvious match, and just because they are both methodological alternatives to conceptual analysis does not mean they are necessarily compatible. After all, CE is still concept-driven, while X-Phi, at least in its infancy, claimed to reject concept-driven investigations. 11 Moreover, X-Phi does not seem to have the normative aspect characteristic of CE, which has been a source of criticism from the opponents of X-Phi, especially for the applied areas of analytic philosophy such as aesthetics.
However, this thought is countered by Andow (2020) who points out that "although normative questions aren't settled by descriptive answers, answering normative questions frequently requires a lot of descriptive information. There is lots of descriptive information that the experimental philosopher could gather which would be relevant in normative debates about concepts" (2020, p. 4). Andow then goes on to make the case that CE could be a fully experimental project, which is more than what I will attempt in this paper. But at the very least it indicates that he is not particularly worried about the compatibility of the two methodologies, and I agree with him.
In relation to the worry about the lack of normativity of X-Phi, Fisher has pointed out that experimental philosophers should be more pragmatic in their examination of current philosophical practices. In discussing the methodological goals of philosophy, Fisher points to 'Pragmatic Conceptual Analysis', which he describes as "a methodology that focuses not on what has been causing us to apply our concepts, but instead upon the beneficial effects that our concept usage has regularly caused" (2015, p. 415). This is not quite the same as CE as I have described it because it does not seem to have the revisionist element characteristic of CE, but it definitely shares its pragmatic concern. And just like Andow, Fisher is optimistic about the compatibility, and more importantly the potential of X-Phi and pragmatic projects. He goes on to argue that X-Phi could play a discovery role in such projects: it could be used to "discover how our concept usage regularly yields beneficial outcomes, so that we can then craft philosophical analyses that will enable us to yield such beneficial outcomes more consistently" (2015, p. 427).
I shall revisit the idea that X-Phi is not sufficiently normative in Sect. 3, as it could still have some bite when applied to aesthetics. So at this point, the only compatibility worry that remains is that X-Phi is not conceptual enough. But this worry is misguided. The fact is that X-Phi does deal with concepts. Just like X-Phi in epistemology deals with knowledge and belief, 12 both the philosophical concepts investigated 11 By 'concept-driven' I mean something like what Knobe and Nichols (2007) were talking about when describing conceptual analysis, i.e. the 'analysis of one concept in terms of another'. As they stated then, the aim of X-Phi is rather 'to provide an account of the factors that influence applications of a concept, and in particular, the internal psychological processes that underlie such applications ' (2007, p. 5), that is, the phenomena surrounding the concept, rather than the concept itself. 12 A clear instance of X-Phi focusing on concepts is the 'Geography of Philosophy Project', which states on its website: 'the question that motivates our project is whether fundamental philosophical concepts […] are religious and cross-cultural universals that are used by people around the world, or if there is from the armchair and the things that play a role in everyday interactions among the folk, X-Phi in aesthetic can and does deal with things like art, beauty, aesthetic experience, aesthetic judgement, taste, and other concepts of that sort.
Moreover, conceptual engineers do not by any means dismiss phenomena in favour of concepts. Again, this probably depends on how conceptual engineers understand concepts, but the case can be made that the relation between concepts and phenomena is an intimate one. Studying concepts does not exclude studying phenomena and vice versa. This is something that conceptual engineers like Haslanger understand well. Her project to ameliorate the concept of woman stems from observing gender in the world and seeing the oppression embedded in it. Her solution is to change the concept of woman in the hope that it will in turn change the reality of being a woman.
Finally, another way of pushing X-Phi and CE closer together might be to look analogically at their respective attitudes towards traditional analytic methods. On the one hand, CE seeks to find deficiencies in our concepts. On the other, X-Phi has made some of its most important contributions to analytic philosophy (through what was called the 'negative programme') by showing that armchair philosophers regularly use questionable, and sometimes even bad evidence in order to theorize about our concepts. In both projects, we find a debunking stance towards traditional philosophical methods. Conceptual engineers then take an ameliorative stance, to which we can perhaps compare the 'positive programme' of X-Phi. In my opinion, this is a little bit less tangible given that experimental philosophers have not yet explored the full potential of positive X-Phi. However, I still think this shows that X-Phi and CE have more in common than we might think, and that there is nothing in principle which would prevent X-Phi and CE from working together. Besides, the compatibility of both projects might become clearer once I show in more detail how each can benefit the other.

Is There Such as Thing as CE in Aesthetics?
Before sketching the role X-Phi could play in setting up CE in aesthetics, we must examine whether it is truly the case that CE currently does not exist in the field. After all, conceptual engineers have in the past identified projects, which predated the birth of the so-called CE movement and yet had all the characteristics of engineering projects. One instance of this is Clark and Chalmers (1998). They offer a famous account of cognition as 'extended', which leads to a reconfiguration of the concept of belief. As Cappelen (2018) notes, their "goal is not to describe our current concept of 'belief'-they want us to revise our current concept" (2018, p. 11) Footnote 12 (continued) diversity in these concepts. Our focus is on the following concepts: understanding, knowledge, and wisdom'.
to include beliefs formed with the help of external devices, like Otto's notebook in their thought-experiment. 13 And while they did not claim to be doing conceptual engineering-and could not really have done so, seeing as the field was not yet developed-Clark and Chalmers did in fact identify a deficiency in a key concept in philosophy of mind and offered an improved version.
So perhaps there are similar cases of 'hidden' CE in aesthetics waiting to be uncovered. One candidate I wish to put forward is Irvin's work on the notion of aesthetic experience (Irvin, 2008a(Irvin, , 2008b(Irvin, , 2014. She has argued on multiple occasions that certain everyday experiences, while lacking the characteristics traditionally required by the concept of aesthetic experience, should count as instances of aesthetic experience. The deficiencies she highlights in the traditional concept are of both kinds identified by conceptual engineers, i.e. practical and theoretical. First, she claims that the traditional concept of aesthetic experience is non-natural, which is a type of conceptual deficiency insofar as good concepts should 'carve nature at its joints'. 14 And so, "if an account of aesthetic experience or appreciation picks out phenomena that turn out not to be empirically viable, we should conclude that it has picked out the wrong things" (Irvin, 2014, p. 45).
Second, she claims that "aesthetic experience and appreciation should not turn out to be rare, exalted states accessible only to art experts or the preternaturally gifted" (Irvin, 2014, p. 45). Here, it seems that she is talking about a practical deficiency with social consequences. The fact that the concept of aesthetic experience does not accommodate 'everyday' experiences like "admiring a garden in full bloom, having one's breath taken away by the beauty of one's sleeping child, and listening with enjoyment to music on the radio" (Irvin, 2014, p. 45) is exclusionary. Experiences which a lot of people would deem very meaningful are not considered 'aesthetic' experiences, and therefore are not valued as such.
Finally, I also believe that one of Irvin's strategies to improve the concept of aesthetic experience is similar to one that conceptual engineers themselves use in some cases. They call this strategy 'splitting', as it amounts to splitting superimposed concepts into more distinct and specific concepts. Irvin proposes to split the current concept of aesthetic experience into three concepts: (1) aesthetic experience, (2) aesthetic appreciation and (3) deep aesthetic appreciation. However, I do not think she uses this strategy because she believes that the traditional concept of aesthetic experience is superimposed. Rather, she wishes to include a type of experiences, i.e. 'everyday' experiences, into the concept and to do so she resorts to simplifying the concept of aesthetic experience and creating two other concepts, aesthetic appreciation and deep aesthetic appreciation, which will then satisfy the more complex aspects that used to be tracked by the original concept. And so, while Irvin did not claim to be engaged in CE, I think it is possible to look at her work on aesthetic 14 Here I am still following the deficiencies delineated by the ConceptLab. 13 In the thought experiment, Otto and Inga both wish to go to the museum. Inga remembers the location of the museum from past visits and heads off. Otto, however, has Alzheimer's and uses a notebook to recall the address of the museum. The thought experiment is designed to show how external memory devices, like Otto's notebook, can play the same functional role as 'traditional' memory. experience through that lens, and perhaps doing so could give us new insight on her proposal. 15

So CE in Aesthetics Might Already be Occurring-How can X-Phi Help?
Now it is time to turn to X-Phi and examine the role it can play in setting up an explicit CE project in aesthetics that appeals to the existing resources in the CE literature. As I see it, the main contribution X-Phi can make to get a CE project up and running in aesthetics is by identifying deficient concepts. More specifically, X-Phi can identify deficient concepts but also, in some cases, the type of deficiency a given concept might have. This is an important distinction: for instance, some X-Phi projects might be able to tell us that there is something wrong with a concept because, in its current form, it does not seem to be shared outside of philosophy.
An example of this in aesthetics is the series of studies led by Cova and Pain (2012) and later Cova (2019) on aesthetic judgement. It is fair to say that aesthetic judgment is one of the central concepts of aesthetics and, as Cova notes, intersubjective validity has for a long time been its main characteristic. That is, aesthetic judgements are often taken to be both subjective and right or wrong. However, his studies have shown that the folk, even across different demographics, do not take their aesthetic judgements to have intersubjective validity. 16 They simply do not apply the concept of aesthetic judgement the same way that philosophers do. As such, the concept of aesthetic judgement, used and theorised upon by philosophers over the 15 At this point, other candidates for CE in aesthetics might have come to the reader's mind. For instance, one might think of Walton's representation as make-believe theory (1990). It might very well be the case that many more theories and projects fit the bill, but before jumping to conclusions, it is crucial to remember that CE is not simply about the intention to change how we conceive of a certain notion or concept. First, this intention must be reasoned. Irvin's reasons might be about the inclusivity and universality of aesthetic experience, for instance. Then, there is a specific process that must be followed, as detailed in the case of Irvin. So, while I am perfectly ready to accept many more candidates for CE in aesthetics, I believe we must study these candidates carefully before claiming the ubiquity of CE in the field. 16 Arguably, there are some flaws with the Cova and Pain studies, as with many of the early X-Phi studies in any philosophical fields. For instance, for Cova and Pain, a participant is marked as rejecting intersubjective validity if, when confronted with a vignette presenting an aesthetic disagreement between two individuals, they choose the option that claims 'Neither is right or wrong. It makes no sense to speak in terms of correctness in this situation. Everyone is entitled to his own opinion'. It has been pointed out to me by an anonymous referee on a previous draft of this manuscript that the last sentence is especially problematic as it amounts to a truism, which would make it an appealing option, and something that both the aesthetic realist and anti-realist could accept. I do agree that this option is not ideally phrased to tease out a real denial of intersubjective validity. However, in the context of the full survey, which includes three other options [(1) one of parties to the disagreement is right and the other is wrong, (2) both are right, or 3. both are wrong], choosing the 'truism' option is not meaningless. It implies that the participant denies aesthetic realism (or normativism, as Cova and Pain put it), which was option 1. They also deny antirealist relativism, i.e. option 2 and 3. This does not leave much logical space. What is left is an antirealist position that also denies the possibility of aesthetic judgements having any truth-value-a position Cova and Pain refer to as expressivism-but that also amounts to the denial of intersubjective validity. So, even if the phrasing of option 4 is misleading, I do not believe that it invalidates the conclusion of the study. years, might very well be defective in the sense that it does not represent the relevant phenomenon. 17 In my view, this would already make a fine contribution, given that at least half of the CE project is dedicated to finding deficient concepts (the rest being dedicated to fixing them, and perhaps putting them back into circulation), and a strong reason for why conceptual engineers should enlist the help of experimental philosophers. Specifically in aesthetics, experimental philosophers have already identified a few deficient concepts, so in that sense conceptual engineers already have somewhere to start. But I also believe that, sometimes, X-Phi can do even better. Once conceptual engineers have been alerted to a deficient concept, they also need to know the type of deficiency they are dealing with in order to know what kind of fix is required.
The issue with X-Phi surrounding intuitions, very broadly construed, is that it only signals the deficiency but does not analyse it. 18 If there is wide disagreement in the intuitions generated by a given concept, thereby indicating that the concept might be deficient, there must be a reason why. In the case of aesthetic judgement presented above, maybe the deficiency is due to some intra-and interpersonal effects. This type of effect, according to conceptual engineers, occurs when our concepts influence our relations to ourselves and others in a negative way. Perhaps, conceptualizing aesthetic judgements as having intersubjective validity would lead us to assert that our own aesthetic sensibilities are superior to someone else's if they were to disagree with us, and this is something the folk want to resist. This would make the concept of aesthetic judgement practically deficient.
However, with X-Phi surrounding intuitions, we can only speculate about the type of deficiency we are dealing with. And if you recall, identifying the deficiency is only the beginning of the CE process, which is then followed by fixing the deficient concept and applying the revised concept. In other words, this kind of X-Phi will not yield much progress for CE in aesthetics, given that all it can do is signal a conceptual deficiency. Granted, identification of the deficiency is perhaps a burden that falls on the shoulders of conceptual engineers and not on those of experimental philosophers. But I think that in some cases, X-Phi can, and in fact should, go further.
For instance, there is a cluster of experimental work on aesthetic communication (Liao & Meskin, 2017;Liao et al., 2016;Andow, 2019;Meskin et al., in preparation). These studies are attempting to address a puzzling issue with aesthetic communication, namely the fact that it does not seem to behave like non-aesthetic communication. Liao and Meskin (2017) and Liao et al. (2016) showed that aesthetic 17 Critics of X-Phi have argued many times that disagreement between the folk and philosophers is not problematic for philosophy. After all, the philosophers might just be right and the folk wrong. I do not think that these arguments are particularly strong, as do many experimental philosophers and methodologists, including Alexander (2010) and Weinberg et al. (2010). Moreover, if the implicit or explicit aim of theorizing about aesthetic judgement is to capture the relevant phenomenon, as it applies to both philosophers and non-philosophers, then this is sufficient to claim that the philosophical concept of aesthetic judgement is deficient. 18 By 'X-Phi surrounding intuitions' I am referring to studies that target intuitional evidence in philosophy. These studies do so with certain identifiable methods, which include vignettes and surveys. Cova and Pain's (2012) and Cova (2019) studies are among them. adjectives are not used in the same ways as their non-aesthetic counterparts and have offered various explanations. Moreover, Andow (2019) and Meskin et al. (in preparation) have confirmed the thought common in the literature that aesthetic testimony does not have the same epistemic standing as non-aesthetic testimony and proposed an explanation for this asymmetry.
While I think that these studies have identified certain important conceptual deficiencies, I am not entirely sure there is one concept they are all targeting. Rather, it seems like they have identified a number of discrepancies and deficiencies in an overarching theory of aesthetic communication, which involves concepts such as aesthetic testimony, knowledge, judgement and language. 19 Perhaps, all of these concepts are deficient in their own way. But the greater contribution of these studies is that they also attempt to pinpoint the source of these discrepancies and deficiencies. This information, which is obtained empirically, will almost certainly be of great value to a conceptual engineer wanting to ameliorate the concepts at play in a theory of aesthetic communication.
I will focus on one example. In his (2019) studies on aesthetic testimony, Andow tested different explanations for the epistemic asymmetry between aesthetic and non-aesthetic testimony. In study 2, the hypothesis tested is that the asymmetry could be explained by non-normativists beliefs and attitudes towards the aesthetic. Simply put, if people think that aesthetic matters are ultimately relative or subjective (i.e. non-normative), then they should also think that aesthetic testimony is of lesser epistemic value than non-aesthetic testimony because we cannot directly transmit relative or subjective information to others. This hypothesis seems to be pointing at a theoretical deficiency in the relevant concepts of aesthetic testimony and aesthetic knowledge: if the concept of aesthetic testimony implies both subject-relativity (seemingly a characteristic of the concept of the aesthetic) and transferability (supposedly a necessary characteristic of the concept of testimony), then it is combining inconsistent concepts. As it turns out, Andow's study did not find the expected effect and concludes that until further research is done on this particular hypothesis, we must look for another explanation.
The one Andow settled on is quite interesting. In study 3, he hypothesised that the asymmetry could be due to a certain interpersonal effect, specifically, the perception that aesthetic matters are subject to higher levels of disagreement and deception. In other words, people do not believe or trust each other to be epistemic authorities regarding aesthetic matters and thus take aesthetic testimony to be of lesser epistemic value than non-aesthetic testimony. If this is the case, then the conceptual deficiency we are looking at is practical in nature. Andow's study 3, which confirmed once again the asymmetry between aesthetic and non-aesthetic testimony, also established that, to some extent, this interpersonal effect of disagreement and deception does explain the asymmetry. Interestingly, Andow suggested that "the explanation for the asymmetry is a least in part due not to a principled distinction between the aesthetic and nonaesthetic domains but rather to this more mundane difference" (Andow, 2019, p. 55). By "principled distinction", I think that he is in fact pointing at a more theoretically grounded explanation, whereas the "more mundane difference" is more practical in nature. Of course, further work could be done to find the exact source of this practical deficiency, which in turn could support an ameliorative project even more, but for now, it is sufficient that I have shown, via a specific example, that X-Phi in aesthetic has arguably signalled and, more importantly, has identified specific conceptual deficiencies in aesthetic concepts and as such should inform any related CE project in the field.

Fixing Aesthetic Concepts and the Purpose of Aesthetic Concepts
Given the conclusion of 3.2, it also makes sense to ask whether X-Phi is capable of aiding in the fixing of deficient concepts. This is a more difficult question. First of all, if X-Phi is capable of identifying conceptual deficiencies as well as fixing them, we might not need CE in aesthetics (or any other field where X-Phi has been developed). It seems quite implausible that both projects would have been developed simultaneously, without experimental philosophers and conceptual engineers noticing that they were effectively doing the same thing.
However, this is perhaps a caricatural (and ultimately anti-pluralistic) conception of philosophical methodologies in practice. As I mentioned before, Andow (2020) argues that CE might be fully experimental. He makes this argument by describing a hypothetical case of CE and showing how every stage can involve empirical data. In particular, he argues that even at the implementation stage of CE, i.e. putting the revised concept back into circulation, which I have mostly set aside in this paper, CE would benefit from empirical information. Experimental philosophers could for instance test different kinds of implementation strategies and assess their effectiveness by measuring the uptake of the revised concept in the relevant population. I am broadly sympathetic to this thought, but do not wish to make any further speculation, especially in the case of aesthetics given the current state of CE in the field.
Moreover, as I stated when discussing the compatibility of X-Phi and CE in Sect. 2, it seems that X-Phi still lacks the normative power that CE has to ameliorate concepts. So X-Phi's contribution, even if present at every stage, is meant to be only secondary, providing all the relevant information conceptual engineers might need to then make normative decisions. More specifically, I believe that the ameliorative potential of X-Phi comes down to our interpretation of the 'ought' of CE. 20 For instance, if the 'ought' is theoretical, that is if a concept ought to serve some theoretical purpose that philosophers have, e.g. provide good explanations, then perhaps X-Phi can help more directly. But if we have certain practical purposes, e.g. eradicate the linguistic effects that participate in the oppression of women, then maybe not so much.
So, to return to our subfield, what is the purpose of aesthetic concepts? Is it a theoretical purpose, e.g. do we wish to describe and explain aesthetic phenomena, or is it a practical purpose, e.g. do we want to say something about what kind of aesthetic appreciation we should develop, whether it should be inclusive or exclusive, etc.? Take the concept of taste: should aestheticians try to describe what taste is in the world, how we acquire it and use it? Or should they try to say something about what taste should be like (e.g. selective or indiscriminate) and how it should be cultivated? Think for instance of certain projects in 'everyday' or 'social aesthetics that offer strategies to overcome discriminatory taste for thin, conventionally beautiful bodies. 21 While I think there is space in aesthetics for projects on both ends of this spectrum, until we state the purpose of the conceptual work done in any given project, we cannot know the full extent of the contribution X-Phi can make to CE. Nevertheless, it seems highly implausible that no contribution would be made at all.

How Would CE in Aesthetics Benefit the X-Phi Project?
I think it is clear at this point that CE is capable of benefitting from turning to experimental philosophers for support, in aesthetics particularly. However, I also think that experimental philosophers have something to gain in this partnership as well. I will suggest two main reasons for why this is the case.
First of all, as any experimental philosopher knows, experimental resources, i.e. the time, effort and funding it takes to offer meaningful empirical information, are scarce. Moreover, there is pressure, coming mostly from armchair philosophy, to prove that X-Phi can yield philosophical progress. To put it crudely, X-Phi has to be cost-effective and experimental philosophers who wish to actively participate in a philosophical community must keep this in mind. On top of designing scientifically sound studies, this means targeting research areas that are (or can be) of interest to armchair philosophers. While this attitude alone cannot ensure that there will be a real collaboration between experimental and armchair philosophers, it is a step in the right direction. By supporting a CE project, in aesthetics for instance but in other areas as well, experimental philosophers can take this step. CE, while also being an alternative metaphilosophical project, remains closer to armchair philosophy in the sense that conceptual engineers often look to armchair philosophy for deficient concepts to ameliorate, with the aim of putting the ameliorated concepts back into circulation within armchair philosophy. 2223 And it seems that conceptual engineers are conscious of prioritizing concepts, which are so central to either our philosophical or our everyday theoretical frameworks that any conceptual deficiency can have truly damaging effects (think about some of the projects I have mentioned on the concepts of truth or woman). If X-Phi can have a strategical role to play in this process, then it can remain in the philosophical loop, so to speak, thereby making efficient use of its resources. This is not to say that X-Phi does not yield philosophical progress on its own. Many have argued that it does. 24 Rather, the point is that to convince the doubters that X-Phi can yield progress, teaming up with CE is one way to do so.
Another criticism regularly addressed to experimental philosophers, and to anyone who wishes to use empirical methods to supplement philosophical claims, is that X-Phi is irrelevant when it comes to making normative claims. Naturally, this criticism mostly affects the areas of philosophy where normative claims are a regular part of the repertoire, such as aesthetics. Among the critics, Dorsch (2014) has argued that when building a theory of aesthetic appreciation and judgement, which should include an explanation of how we pick up on aesthetic properties, empirical evidence is methodologically useless. This is because things like aesthetic properties are "typically normative: they are either values, or by default linked to values in such a way as to constitute reasons for evaluation" (2014, p. 78), and we cannot have direct empirical access to normative kinds. 25 This type of objection is not only common, it is also hard to shake off because it rests on deep-seated assumptions about the concept of the aesthetic, the nature of aesthetic properties, and the rightful subject matter of philosophical aesthetics. However, if X-Phi can actively support a CE project in aesthetics, then it can play an important methodological role in a project that is ultimately normative. Admittedly, this not a direct response to the objection, but a way of pushing back against the attitude that X-Phi and other empirical methods cannot be relevant to philosophy on all levels, descriptive or normative. Some of the assumptions that lead to this type of objection might still need to be revised at some point, but for now, we can rest assured of the fact that X-Phi has something significant to offer to normative projects in philosophy. 22 An objection to this, called the Discontinuity Objection, argues that the 'amelioration' or changes that the concepts undergo actually amounts to a change in subject. For an overview and response, see Prinzing (2017). 23 Some conceptual engineers might object to the idea that CE is closer to armchair philosophy than X-Phi. More specifically, I am thinking that Cappelen, who on the one hand objects to X-Phi on the grounds that armchair philosophy does not rely on intuitions (2012), and on the other argues that the aim of CE is to change the world via our concepts (2018). 24 See again, Alexander (2010), and for aesthetics specifically, Torregrossa (2020). 25 Another instance of this type of objection is Zangwill (2019).

Conclusion
I have defended three claims in this paper. First, I have argued that X-Phi and CE are two compatible metaphilosophical projects, and if CE were to be officially developed in aesthetics it would be compatible with current experimental aesthetics. Second, I believe there are good reasons for experimental aestheticians to encourage the development of CE in their field. For one, the way X-Phi is currently practiced in aesthetics makes it capable of contributing significantly to CE. Not only is X-Phi in aesthetics capable of identifying deficient concepts, but it is also capable of identifying the type of deficiency a given concept is dealing with. This should make CE in aesthetics considerably more effective. Moreover, experimental aestheticians have something to gain in supporting the development of CE in their field. As far as I can see, the only things that might be standing in the way of X-Phi in aesthetics are practical issues, such as the lack of resources and perhaps the lack of wider appreciation of X-Phi results in aesthetics. Supporting another metaphilosophical project with normative potential such as CE will almost certainly help remedy some of these practical issues. On a more hypothetical note, I am tempted to think that when alternative methodological projects support each other successfully, it plays in favour of a form of methodological pluralism, which in turn might promote philosophical progress.