British Undergraduate Review 1(1) 2020 by Arden Foster-Spink

British Undergraduate Philosophy Review Vol. 1 (2020)

Editor-in-chief: Alexander Arridge Mansfield College, Oxford

British Undergraduate Philosophy Review www.britishundergraduatereview.org/philosophy/ All rights reserved. ÂŠ2020 British Undergraduate Philosophy Review, and the authors. No part of the British Undergraduate Philosophy Review may be reproduced or used in any form or by any means (graphic, electronic, or mechanical, including photocopying, taping, recording, or by any other information storage and retrieval system) without written permission from the British Undergraduate Philosophy Review editorial team, and the authors.

British Undergraduate Philosophy Review vol. 1 (2020)

Imprint Editor-in-Chief: Commissioning Editor: Typesetting Editor: Chair (Brit. Undergrad. Rev.):

Alexander Arridge Helena Lang Matthew W. G. McClure Arden Foster-Spink

Reviewers: Matthew W. G. McClure Martin Walter Niederl Khai Ojehomon Harry Pendlebury Christian Pitt Amedeo Robiolio Jack Sagar Alex Stanley Josh Tulloch Marlene Valek

Martina Azlor Rose Bache Theo Beck Alex Blanc Sam Bowerman Harry Croxford Jules Desai Charlie Gorst Jack Graveney Jay Hollis Alexia Leclerc Contributors: Guillaume Andrieux George Campkin Rebecca Clark Alexander Gruen Isaac Hadfield Alexander Houghton

Milena Koot Crystal Lam Lorenz von LĂźpke Viola Nassi Willa Saadat Jared Smith

Contents Defusing the Present Danger of the Future Retrospective Excuse Guillaume Andrieux

Do Democratic Citizens Have a Duty to Vote? George Campkin

What is MacKinnon’s Argument Against Pornography? Rebecca Clark

Phenomenal Overﬂow: What an Auditory Analogue of Sperling Does Show Alexander Gruen

‘Ought Implies Can’ and the Possibility of Group Obligations Isaac Hadﬁeld

A Defence of Hilary Putnam’s Theory of Reference via Temporal Externalism Alexander Houghton

A Brief Conceptual Analysis of what ‘Implicit Attitude’ Means Milena Koot

A Logical Challenge to the Future-like-ours Argument and Illegalizing Abortion Crystal Lam

Hermeneutical Injustice, Locutionary Silencing, and Bad Sex Lorenz von Lüpke

Testimonial Injustice in Virtue of Due Credibility Viola Nassi

A Case Against Dreaming as Hallucination Willa Saadat

103

Recognition in Hegel’s Phenomenology Jared Smith

109

Editorial Welcome to the first edition of the British Undergraduate Philosophy Review, a new journal showcasing the very best that undergraduate philosophy has to offer. The papers making up this journal cover a vast range of topics: some grapple with problems as old as the discipline of philosophy itself, while others make important contributions to subjects only recently coming to prominence in philosophy. I’m sure that readers from across the philosophical spectrum will find something of interest to them in these pages. There are papers demonstrating the practical impact, and importance, of good philosophy. Rebecca Clark argues powerfully for a shift in how we conceive the harms of pornography. Viola Nassi argues that one can be the victim of testimonial injustice even if one is afforded the appropriate epistemic credibility, hence enhancing our understanding of the harms done in phenomena such as mansplaining. Lorenz von Lüpke applies the notions of hermeneutical injustice and locutionary silencing to deepen our understanding of valid sexual consent: they argue that the validity of an act of consent can be undermined if it takes place under the influence of an oppressive social context. And writing as America goes to the polls, I am reminded of George Campkin’s paper detailing a framework for explaining democratic citizens’ duty to vote, and Isaac Hadfield’s paper arguing that groups failing to qualify as agents can nonetheless have abilities and thus be subject to obligations. There are papers making powerful use of empirical evidence to substantiate or motivate their philosophical claims. Alexander Gruen defends the thesis that phenomenal consciousness overflows access consciousness on the basis of evidence from an auditory analogue of Sperling’s famous partial report experiment. Willa Saadat makes use of brain imaging data to inform their consideration of one of the most mysterious yet beautiful human experiences: dreaming. This data is used to argue against the philosophical orthodoxy that dreams are hallucinations. There are papers contributing to longstanding debates across philosophy. Guillaume Andrieux defends the not uncontroversial position that the truth of free will scepticism neither entails that all actions are permissible, nor threatens our ordinary practice of holding agents morally responsible for their actions. Crystal Lam offers a rigorous dismantling of the Future-like-ours argument for the impermissibility of abortion, and Alexander Houghton argues that Temporal Externalism (the view that the meaning of words at a given time can be determined by the way those words are used after that time) supports Putnam’s theory of reference regarding ordinary natural kind terms. A thorough analysis of recognition in Hegel’s Phenomenology is pro-

vided by Jared Smith, while Milena Koot engages with the contemporary literature on ‘implicit attitudes’ to argue that they should be analysed as an ontologically diverse range of mental processes operating under suboptimal conditions. Before inviting you to sit down, make a cup of tea, and enjoy the excellent philosophy previewed above, some acknowledgments are in order. For their diligent work I would like to thank our team of reviewers. Special thanks also must go to Helena, Matthew, and Arden for their incredible work facilitating the publication process from start to finish. Lastly, I should note that the BUPR aims to carry the flame first lit many years ago by the British Journal of Undergraduate Philosophy: we hope, as the BJUP did so well, to shine a light on some of the best undergraduate philosophy from the UK and beyond for many years to come. Alexander Arridge Mansﬁeld College, Oxford, England

Defusing the Present Danger of the Future Retrospective Excuse guillaume andrieux

N his Free Will and Illusion, Saul Smilansky (2000) defends an atypical position. He is a free will skeptic, that is, he doubts that we possess libertarian free will (also called contra-causal free will, roughly the ‘ability to do otherwise in the exact same situation’). This position is almost always accompanied by a skepticism toward a certain type of moral responsibility, called basic desert moral responsibility (Shaw et al. 2019), defined as follows:

For an agent to be morally responsible for an action in the basic desert sense is for it to belong to her in such a way that she would deserve blame if she understood that it was morally wrong, and she would deserve credit or perhaps praise if she understood that it was morally exemplary. The desert invoked here is basic in the sense that the agent, to be morally responsible, would deserve the blame or credit just because she has performed the action, given sensitivity to its moral status, and not by virtue of consequentialist or contractualist considerations. (Pereboom 2014) According to Smilansky, if we were to fully internalize the free will skeptic’s theoretical stance, it would create practical difficulties ‘generating acute psychological discomfort for many people and threatening morality — if, that is, we do not have illusion at our disposal’ (Smilansky 2000: 166). Accordingly, he argues for illusionism, the view that the skeptic ought to preserve other people’s intuitive belief in libertarian free will, even if that belief is false. Smilanslky’s illusionism is a complex position that has been challenged by other free will skeptics, usually called optimistic skeptics, who claim that the practical consequences of free will skepticism, and in particular skepticism about basic desert moral British Undergraduate Philosophy Review 1 (2020)

Future Retrospective Excuses

responsibility, are neutral if not preferable compared to those of living with libertarian beliefs (Caruso 2016; Pereboom 2014; Waller 2011). In this discussion I would like to focus on one of Smilansky’s arguments that does not seem to have been directly considered by the optimistic skeptics (except for Harrison (2009), although I will take a route different from his): Present Danger of the Future Retrospective Excuse: the danger that people’s knowing at the time of choice that in the future, when that choice is judged retrospectively, they will have an excuse based upon the hard determinist perspective might affect their present choosing. (Smilansky 2000: 314) If the Present Danger of the Future Retrospective Excuse (henceforth PDFRE) argument is sound, then this will be a strong blow against the sustainability of the free will skepticism perspective in practice (Smilansky 2000, 2012, 2017). Smilansky would then be justified in identifying the Future Retrospective Excuse (henceforth FRE) argument as a potent threat to moral behavior, causing considerable difficulties for social life. Indeed, it is hard to conceive how a society where everyone is justified in using such a moral excuse would function, and it is quite uncontroversial to see such a development as disastrous for our own. If the free will skepticism perspective entails PDFRE, then this would make this view untenable in practice. My aim in the following pages is to show that the skeptic can resist the PDFRE argument. In a first part, I will argue that the FRE argument is unsound while in a second part I will use an account of moral blame that is compatible with free will skepticism to show how the PDFRE argument can be undermined. As a result, I will argue that the danger of the Future Retrospective Excuse can be easily escaped by the skeptic.

1 The issue with the Future Retrospective Excuse Smilansky’s original formulation of PDFRE is aimed at the hard determinist view. While not all free will skeptics are hard determinists (Shaw et al. 2019: 4), all hard determinists are free will skeptics. Since PDFRE hinges on the hard determinist’s rejection of basic desert moral responsibility and that free will skeptics share this rejection on similar grounds (Shaw et al. 2019), the argument applies as well to free will skepticism. We can lay out the PDFRE and FRE arguments as follows.

Guillaume Andrieux

Present Danger of the Future Retrospective Excuse P1: The FRE argument can convince a significant number of free will skeptics that it is morally permissible for them to decide to do any action. P2: If the FRE argument can convince a significant number of free will skeptics that it is morally permissible for them to decide to do any action, then the FRE argument seriously threatens our moral behavior. C: The FRE argument seriously threatens our moral behavior. Future Retrospective Excuse P1: S does not have basic desert moral responsibility. P2: If S does not have basic desert moral responsibility, then S will not be morally responsible in a desert-based sense when S will have Φed. P3: If S will not be morally responsible in a desert- based sense when S will have Φ-ed, then S will not be morally blameworthy in a desertbased sense for having Φ-ed. P4: If S will not be morally blameworthy in a desert-based sense for having Φ-ed, then S is not morally blameworthy in a desert-based sense for deciding to Φ. P5: If S is not morally blameworthy in a desert-based sense for deciding to Φ, then it is morally permissible for S to decide to Φ. C: It is morally permissible for S to decide to Φ. In order to resist the PDFRE argument, the free will skeptic might attempt to refute FRE. P1 is one of the core commitments of free will skepticism, and P2 to P4 should also be endorsed by the free will skeptic. P5 may be true in many cases: we can think of a variety of morally neutral or good actions for which P5 holds. However, there are cases where P5 is false, and the task of the free will skeptic is to show that PDFRE involves such cases. To begin, we need to identify the type of case where P5 is false: cases where (a) S is not morally blameworthy in the basic desert sense for deciding to Φ and (b) it is not morally permissible for S to Φ. A classic example would be a serial killer whose mental

Future Retrospective Excuses

illness is a determining cause of his deeds (that is, without it and all other things being equal, he would not have killed his victims). In this case, it is generally accepted that we canâ&#x20AC;&#x2122;t morally blame him in the desert-based sense. Even within a compatibilist account of free will (and the free will skeptic would agree) it will be recognized that the killer effectively lacked the kind of control required to have basic desert moral responsibility for these acts. Nevertheless, we can also recognize that his actions were morally wrong and thus not morally permissible. Indeed, and this is a crucial point, free will skeptics (Haji 1998; Pereboom 2001, 2014) have argued that free will skepticism does not threaten axiological judgement of goodness and badness as they donâ&#x20AC;&#x2122;t necessarily require blameworthiness and praiseworthiness. Condition (a) is satisfied when the agent lacks basic desert moral responsibility, and condition (b) is satisfied when the agent is engaging in a morally wrong act. For the free will skeptic, everyone always lacks basic desert moral responsibility, so agents considering FRE necessarily satisfy condition (a). Concerning condition (b), we can observe that virtually any agent using the FRE argument believes that the action she is considering is morally wrong: if S uses FRE to justify an action that she thinks is morally permissible, then we might wonder why S is engaging in FRE in the first place. Since escaping oneâ&#x20AC;&#x2122;s moral conscience seems to be the only reason why someone might have recourse to FRE, we can justifiably think that such cases are virtually nonexistent and thus irrelevant for the PDFRE argument. When a free will skeptic engages in FRE she might not know that the action she is considering is morally wrong: she believes it but, in some cases, may be mistaken. However, such cases are also irrelevant in the context of PDFRE. If S uses FRE to justify what is in fact not a morally wrong action, then this will be a case of benign FRE that will not threaten moral behavior since the action would not in fact be morally wrong. As a result, cases where S believes that the action is not morally wrong and cases where S mistakenly believes that the action is morally wrong are almost never going to be cases where FRE represents a threat to moral behavior. It follows that the only type of FRE instances at play in the PDFRE argument are cases where S believes that the considered action is morally wrong and that this action is actually morally wrong. All these cases satisfy condition (b). So we have seen that virtually all instances of FRE relevant for the PDFRE argument satisfy both conditions (a) and (b). From that we can conclude that for virtually all instances of FRE relevant for the PDFRE argument, P5 is false. The mistake made by the agent who accepts the FRE argument is that she believes that desert-based blameworthiness and doing a morally wrong action are directly linked, so that it is im-

Guillaume Andrieux

possible for the latter to obtain if the former does not. As we saw above, free will skeptics have argued that this is false. To them, the FRE argument is therefore unsound.

2 Resisting PDFRE with forward-looking blame Unfortunately for the free will skeptic, PDFRE might nevertheless remain sound. The reason for this is that PDFRE’s P1 does not require the soundness of FRE, but merely that it can be accepted by a significant number of free will skeptics. The hypothetical environment that Smilansky sets out for PDFRE is a society where a significant portion, if not majority, of the population espouses hard determinism (in particular skepticism about libertarian free will and basic desert moral responsibility). This can be expected to happen if the views of the hard determinists, and more generally of the free will skeptics, become more visible and convincing in non-academic discourse (avoiding this eventuality is one of the reasons behind Smilansky’s illusionist stance). If this happens, the majority of the free will skeptics will not be academics nor philosophy students and we can reasonably suppose that many of them will not engage in the free will debate in an in-depth fashion, for lack of interest, lack of accessibility or some other reason. Hence there will be an important risk that enough people miss the flaw of the FRE argument (which is not obvious from the lay perspective) and end up convinced by it, making P1 true. An additional worry, perhaps more serious, concerns the ability of the free will skeptic to respond to a wrongful act committed by a fellow skeptic on the basis of the latter’s acceptance of FRE. Let’s imagine the following case. Will and Lena are both free will skeptics (in particular they share the same skepticism about basic desert moral responsibility). Will happens to discover the FRE argument by reading an excerpt from a book, and genuinely believes that it is sound (perhaps because of careless reasoning). Later that day, he is hungry and notices that his colleague Lena has left her croissant unattended on her desk. Recalling the FRE argument, Will concludes that it is permissible for him to eat the croissant and proceeds to do just that. Later, Lena arrives and asks him if he saw anyone taking her missing pastry, to which Will readily answers that it was him. She understandably becomes upset, but he quickly adds that his action was morally permissible, telling her about the FRE argument. Lena knows that FRE is flawed, yet as a free will skeptic she also believes that Will is not responsible in a basic desert sense for believing that the argument is sound, nor for eating her croissant. It then seems that she can’t blame him for what he did, and most importantly, that she will not be able to blame him if he reiterates (since he never has basic

Future Retrospective Excuses

desert moral responsibility). Here we see that PDFRE’s P2 is apparently verified: the acceptance of FRE poses a threat to the moral behavior of the free will skeptic. At this point, Lena’s position appears untenable: the only way for her to preserve moral behavior would be to blame Will for his deed, relinquishing her stance on free will for (at least) a compatibilist account that allows for Will to be blamed in the basic desert sense. But there is at least one way she could preserve moral behavior while remaining a free will skeptic: she could use a notion of blame that is independent from basic desert moral responsibility. Derk Pereboom (2014) offers an account of a ’forward-looking’ morality with an associated forward-looking notion of blame that might well be sufficient to resolve Lena’s problem. Pereboom modifies McKenna’s conversational theory of moral responsibility (Pereboom 2014: 132), grounding it in ‘three non-desert invoking moral desiderata: protection of potential victims, reconciliation to relationships both personal and with the moral community more generally, and moral formation’ (Pereboom 2014: 134). This model functions in three stages. In the first stage (moral contribution), an agent acts such as she ‘at very least implicitly conceives of her action as morally charged and as communicating a particular quality of will’ (Pereboom 2014: 133). This is the case for Will: he has engaged in moral deliberation over his decision of eating the croissant. The second stage (moral address) consists in a request for an excuse or justification from a respondent. Arguably, in our example Lena doesn’t even have to explicitly voice this request since it is implicitly understood by Will. In the third stage (moral account) the agent offers ‘an excuse, a justification or an apology’ (Pereboom 2014: 132). In our case, Will chooses to offer a justification for his act, namely that according to FRE he knew that his action was permissible. In a fourth stage, depending of the agent’s answer and its acceptance by the respondent, the agent may be forgiven or blamed and/or punished. At this point, Lena can explain to Will why FRE is an unsound argument, hence that his justification is flawed. Suppose that Will, who genuinely believed that FRE was sound, understands his mistake. He is blameworthy under the conversational model because this blame is instrumental in realizing the aforementioned moral desiderata. It promotes the protection of victims (he will be less likely to steal other’s pastries in the future), the reconciliation in personal relationships (with Lena) and with the moral community (his other colleagues, friends, etc. who might hear about the incident), and his moral formation (he will not make the careless mistake of accepting FRE in the future). This form of blame is ’forward-looking’ because ‘its aims are future protection, future reconciliation, and future moral formation’ (Pereboom 2014:

Guillaume Andrieux

135). It is therefore distinct from desert-based forms of blame in which the agent is blameworthy just because he has performed the action, regardless of other considerations. This is why this account of blame is available to the free will skeptic. In our example, Will genuinely believed that FRE was sound. But we can imagine an analogous situation with Jim instead of Will, in which Jim knows or suspects that the FRE argument is flawed and yet uses it as a justification for stealing Lena’s croissant. In this situation, Jim acts dishonestly. But since he cannot produce a convincing answer to Lena’s rebuttal of FRE, the blame will remain legitimate. So the forward-looking account of blame remains satisfying in this case as well. We have seen that the free will skeptic has at least one way to respond adequately to a wrongdoer who uses the FRE argument. Let’s now go back to the first worry, namely that there might be a risk that a significant number of people become convinced by the FRE argument if free will skepticism gains in visibility and credence in the public discourse. If this risk is realized, we could perhaps witness a surge of immoral actions due to the moral inhibition that FRE would have on those who accept it. There is a simple step that can be taken to greatly minimize that risk: that the flaws of the FRE argument and the ways to respond to someone using it (and Pereboom’s account of blame may be one candidate among others) be discussed when free will skepticism appears in the public debate. Since academic philosophers are expected to enjoy a role of experts when the free will debate is publicly discussed, they would be the main target for this recommendation. This will have the effects of (1) minimizing the number of people susceptible to fall for the FRE argument and (2) maximizing the number of people competent to act as a successful respondent when FRE is used to justify a wrongdoing (which requires that, like Lena, one knows why FRE is flawed). Thus, if FRE is accepted it will be so only by a minority, which is expected to rapidly dwindle as its members encounter competent respondents. It follows that PDFRE’s P1 will be undermined (as FRE will only be accepted by a minority), as well as P2 (since the action of competent respondents will significantly reduce the threat over time), making the whole argument much less convincing.

3 Conclusion Smilansky’s PDFRE argument threatens to make free will skepticism untenable in practice by identifying in FRE a moral loophole enabling the free will skeptic to believe that any action is morally permissible, with dramatic consequences. However, upon closer inspection it appears that the FRE argument is unsound because one of

Future Retrospective Excuses

its premises is false. This by itself does not do much to reduce the force of PDFRE, but when this point is accompanied by an account of blame that is compatible with free will skepticism (such as Pereboom’s), it provides a solution for the skeptic. If the flaws of FRE and the way to respond to an instance of its use are made sufficiently visible when free will skepticism is discussed outside of academia, then the free will skeptic is on solid ground to resist the Present Danger of the Future Retrospective Excuse. The University, Glasgow, Scotland

References Caruso, G. D. (2016). Free will skepticism and criminal behavior: A public healthquarantine model. Southwest Philosophy Review 32(1), 25–48. Haji, I. (1998). Moral Appraisability: Puzzles, Proposals, and Perplexities. Oxford. Harrison, G. K. (2009). Hooray! We’re not morally responsible! Think 8(23), 87–95. Pereboom, D. (2001). Living Without Free Will. Cambridge. Pereboom, D. (2014). Free Will, Agency, and Meaning in Life. Oxford. Shaw, E., D. Pereboom, and G. D. Caruso (Eds.) (2019). Free Will Skepticism in Law and Society: Challenging Retributive Justice. Cambridge. Smilansky, S. (2000). Free Will and Illusion. Oxford. Smilansky, S. (2012). Free will and moral responsibility: The trap, the appreciation of agency, and the bubble. The Journal of Ethics 16(2), 211–239. Smilansky, S. (2017). Pereboom on punishment: Funishment, innocence, motivation, and other difficulties. Criminal Law and Philosophy 11(3), 591–603. Waller, B. N. (2011). Against Moral Responsibility. MIT Press.

Do Democratic Citizens Have a Duty to Vote? george campkin

1 Introduction OTING is inextricably linked to democracy. It is how representatives are selected and how policies are chosen. Debate over the voting franchise has been central to democratic history the world over. Democratic citizens have the right to vote, but whether they have a corresponding duty to vote is what I will attempt to answer in this essay. After clarifying the question, I will explain what I have termed the ‘Effecting Change Approach’ to establishing a duty to vote. I will then assess whether it is successful and conclude that this approach shows that such a duty can be established, providing some further work is done.

2 Duties and democratic citizens I will first make two clarificatory points about the question, framing the discussion to follow. The first establishes what it is meant by a duty in the context of the question and the second stresses the importance of the application of the question to democratic citizens in particular. 2.1 The duty To have a duty is to ‘be subject to a binding, normative requirement’ (Frazier 1998). The duty-bound person has a moral requirement to satisfy the content of the duty, analogous to a subject who must obey the command of an agent with a right to be obeyed (Frazier 1998). I will differentiate between two varieties of duty to clarify to which type I will refer in this essay. The first is a ‘special obligation’ (Jeske 2019). This type of duty is borne out of a certain situation, whether it is because a promise has been made, or the dutybound person holds a particular role, or that they have made some other commitment which explicitly or implicitly creates the duty. In the case of voting, this might be because the subject promises someone that they will vote in the election and so creates British Undergraduate Philosophy Review 1 (2020)

Democratic Duty to Vote

a duty to vote from making this promise (Lomasky and Brennan 2000). This, however, relies upon an external commitment which need not be present in any more cases than the given example. The duty to vote in such a circumstance is dependent on this unrelated commitment and so does not seem to be a real duty since it trivially derives from this other undertaking inserted into a specific instance. Instead, I will focus on a fuller sense of duty; a generalized, intrinsic type of duty. Its existence and strength does not rely on anything outside of the closed reasoning for the duty since it is independent of external factors. It is this sort of duty that must be established in the context of the question. 2.2 Democratic citizens Now, I will stress the relevance of applying a potential duty to vote to democratic citizens. Democracy, a political system based on collective decisionmaking (Christiano 2018), seems to be most ripe for a duty to vote. Firstly, democratic citizens are afforded the right to vote. Whilst this does not necessarily entail a duty to vote, it opens the possibility. To talk of a duty to vote without first having the right to do so is nonsensical; it would be bizarre to speak of citizens in a dictatorship having a duty to vote when elections are not even held. There are some democratic citizens without this right, such as children and prisoners, but this is not, in itself, philosophically interesting, so I will ignore these groups. It does, however, shine some light on the nature of a potential duty to vote. The right to vote is artificial; it is not a right with which one is born and it can be lost. It is reasonable to assume that a duty to vote would reflect this. A duty to vote is not some sort of natural duty, as refraining from killing conceivably could be. This is why I will focus on a theory seeking to establish the duty to vote through a secondary duty of a more natural sort, grounding the more artificial duty to vote. Secondly, votes cast in democratic operations have significant effect. They decide the government, legislatures and can even make singular decisions on policy. Furthermore, most citizens in contemporary democracies think that there is such a duty (Mackie 2010: 8â&#x20AC;&#x201C;9) and these are the kind of democracies which I will consider; ones in which there are large populations infrequently voting in elections or referenda. I will not be considering the citizens of past democracies like ancient Athens, nor utopian, imagined democracies. The following discussion will be based on the modern reality of democracies and will seek to answer whether their citizens have a duty to vote as framed in this section.

George Campkin

3 Establishing a duty to vote 3.1 The Effecting Change Approach Here, I will set out what I label the ‘Effecting Change Approach’. This route to ground a duty to vote appeals to voting as the way to change the individuals holding political power and alter political policy. This foundation then allows one to suggest there is a duty to vote because of another existing duty which it supports. This secondary duty might be something like a duty to protect oneself or a duty to help others, or even as straightforward as a duty to produce good government (Brennan 2016). The idea is that this secondary duty will entail the duty to vote since it is a component of the broader secondary duty. For example, if there was a duty to help others and one could fulfil this duty by seeking to have implemented policies that are more likely to help others, then it appears that there is a duty to vote as voting is the way to action this fulfilment. Similar lines of reasoning could be used for the other examples above. 3.2 The problem of the Paradox of Nonvoting A problem seems to arise from this approach resting on the assumption that a vote cast by an individual has some effect in bringing about their desired result. By relying on the efficacy of individual votes, it is susceptible to what is termed the ‘Paradox of Nonvoting’: because no single vote has an effect on the outcome of an election, the individual act of voting is instrumentally irrational (Mackie 2010: 2). Take B as the individual’s benefit from winning an election (created by the result supporting some aim of the individual), C as the cost of voting to the individual (mostly opportunity cost arising from time spent deliberating over and then casting the vote) and p as the probability that the individual’s vote causes the winning election outcome. Then, rationally, the individual should only vote when pB − C > 0 (Downs 1957: 246). When the modern democratic context set out in section 2.2 is considered, the probability that the individual’s vote is pivotal in deciding the election is extremely close to zero. This results in the expected benefit (pB) typically being worth far less than a millionth of a penny in monetary terms (Brennan and Lomasky 1993: 119). So, providing that the cost of voting for the individual is higher than this minute value (an evidently sensible assumption), then pB − C < 0, meaning it is irrational for the individual to vote. If this is left to stand, it would undermine this approach. If it is irrational to vote since it has negative expected instrumental value, then the link between the secondary duty and the duty to vote is broken; voting can no longer be taken as a way to fulfil the obligations of the secondary duty and the grounding for a duty to vote is lost.

Democratic Duty to Vote

3.3 Dismantling the paradox There are a variety of potential solutions for the paradox, all at least attempting to show that voting can be rational. The responses can be divided into two sets: those which seek to undermine the above framework of the rationality of voting and those which work within the broad framework. Both have the same end—to demonstrate the possible or definite rationality of voting—but they go about it by different means. A response of the former sort would deny that ‘the sole aim of the voters is to put a candidate into office or a proposal into effect’ (Elster 2007: 217). A prominent example comes from Brennan and Lomasky (1993) in the ‘Expressive Theory of Voting’. It holds that voting is a consumption activity rather than a production activity (Brennan 2016); it is a way of expressing support for the party, candidate, or policy of which one approves, rather than an activity to actually bring about a change in favour of that party, candidate or policy. But, a solution of this sort does not fit the bill for our purposes. If the paradox is dismantled by rejecting that voting is for the purpose of enacting change, the link between a duty to vote and the secondary duty is broken. If voting does not result in action to the benefit of the secondary duty, then a duty to vote cannot stem from that secondary duty. So, whilst a solution of the former sort might stand to provide a rational basis for voting, it cannot solve it in a way that retains the link between the secondary duty and a duty to vote. Instead, the paradox must be dismantled by an approach of the latter sort. Whilst one could resort to relinquishing the importance of being the deciding vote by suggesting, for example, that a voter might be in the ‘causally efficacious set of votes’ (Tuck 2008), the most straightforward solution of this sort is to alter the inputs to pB − C > 0, ensuring that it is satisfied. This could be done by either insisting that the cost of voting (C) is zero, or by holding that the benefit (B) of getting the desired result is very large. Whilst these might be separated into two distinct approaches (Dowding 2005), I would suggest that they can be taken together, as two sides of the same coin. By re-framing the perspective from that of the individual to that of society, one can demonstrate the rationality of voting while retaining the link between the secondary duty and the duty to vote. By moving from an individualistic appraisal of voting to one taking into account the whole political community, one shifts to accounting for everyone’s utility (Lomasky and Brennan 2000: 68). Whilst this might appear just to scale-up the previous problem, it does more than linearly magnify the benefits and costs of voting. The change in degree becomes a change in kind (Lomasky and Brennan 2000: 68). Parfit makes an analogy with moving from someone considering a one-in-a-million chance of killing one person to someone (a nuclear engineer) with a one-in-a-million chance

George Campkin

of killing a million people (Parfit 1984: 74–75). The probabilities are the same for the two individuals, but the stakes are drastically different. Whilst it is irrational to give much thought to the risk in the first instance, it is not in the second one. The stakes in democratic elections are more like that of the second instance; the results will affect the entire population, not just the individual voting. No matter how small the chance of deciding the result, the stakes are just too high to ignore. The benefit input of the equation is therefore huge and so even though p remains the same, pB is amplified significantly. This change in perspective also brings consequences for the costs in the equation. When the frame of reference undergoes this change, the cost of voting diminishes greatly. Spending fifteen minutes to avert a one-in-a-million chance of a nuclear disaster is surely a small price to pay. Likewise, ‘[a]re there really so many more beneficial things one could do with fifteen minutes’ (Barry 1978: 39) than casting a vote with a one-in-a-million chance of improving the situation of the entire political community? This might seem an unwarranted move with the sole purpose of allowing voting to be rational. I would suggest that it is actually a more natural view of the equation of the costs and benefits of voting. To conceive of a duty to vote as ‘primarily self-regarding’ seems odd (Lomasky and Brennan 2000: 67). As indicated in section 2.2, a democratic citizen is part of a collective, bound by shared decision-making. It would seem strange, under these circumstances to take voting as a wholly private and individual action. Furthermore, most of the potential secondary duties for grounding a duty to vote seem to take this perspective. Duties to help others or provide good government are not self- regarding. This should not come as a surprise; it is surely natural for a duty to be in some way selfless, rather than purely prudential. Moral considerations tend not to be focused on fulfilling one’s own wants, but rather on more altruistic ends. For voting to fulfil an obligation of the secondary duty, it makes sense for it to serve the same sense of benefit. 3.4 Tying up loose ends I now have set out how the ‘Effecting Change Approach’ shows that there can be a duty to vote. Two further tasks must be completed before this duty to vote is actually established. The first is to show that voting holds a privileged status in working to fulfil the secondary duty. It cannot just be one possible way in which to work towards fulfilling the obligations of the secondary duty. It must instead be the sole way of effectively fulfilling an obligation of the secondary duty. For the link between the two duties to hold, voting must instrumentally satisfy a component of the secondary duty which

Democratic Duty to Vote

either cannot be done by an alternative method, or at least cannot be done as effectively as by voting. If it could easily be satisfied by other means, then the grounds for a duty to vote are lost as one can refrain from voting and still fulfil the secondary duty. This condition can be obtained fairly straightforwardly since ‘[e]ven though voting is surely not the only way to affect government, it seems to be the only way to choose it’ (Maskivker 2019). Voting holds a special position in democratic society as although citizens can protest, make arguments for their political views, or engage in civil disobedience, none of these actually selects the government in power. They might be able to somehow influence it or perhaps make it more likely that they are replaced, but voting is enshrined in law as the sole method to directly change the government and choose policies. It therefore attains the necessary privileged status in working to effect political change and so the link between the duty to vote and the secondary duty is obtained. This brings me to the second task. Throughout this essay, I have referred to an abstract ‘secondary duty’. I have done this so as not to inadvertently close off options prematurely when assessing the ‘Effecting Change Approach’. But, for the duty to vote to be fully established, the secondary duty must be made concrete. One of the candidates for a secondary duty must be chosen and then shown to be legitimate. If this is not done, then one is guilty of pushing back the question to the secondary duty without answering it. Note, however, that this is not a fault of the sort identified in section 2.1 since the secondary duty on which the duty to vote rests is not trivially inserted into the reasoning; it is instead a generalized duty, intrinsically linked to the duty to vote by the ‘Effecting Change Approach’. So, in fulfilling this final task, one will legitimately and fully ground the duty to vote.

4 Conclusion In this essay, I have shown how a duty to vote can be established. After setting out in what this duty would consist (section 2), I expounded the ‘Effecting Change Approach’ to establishing it (section 3.1). Whilst the Paradox of Nonvoting presented an objection to voting being able to effect change (section 3.2), it was overcome by a justifiable shift in perspective from the individual to society (section 3.3). The link between the duty to vote and the secondary duty was shown to resist a challenge to the status of the action of voting (section 3.4). What I aim to have shown in this essay is that the ‘Effecting Change Approach’ is an unproblematic way in which a duty to vote can be established. Strengthening this to show that the duty is established requires one to prove the secondary duty to be legitimate. This is beyond what reasonably can be

George Campkin

hoped to be achieved in this essay since it requires establishing an even more complex duty. But, this being completed, I hope to have laid the groundwork for stretching this secondary duty into a duty to vote. University of Warwick, Coventry, England

References Barry, B. (1978). Comment. In S. Benn (Ed.), Political Participation: A Discussion of Political Rationality. Australian National University. Brennan, G. and L. Lomasky (1993). Democracy and Decision: The Pure Theory of Electoral Preference. New York: Cambridge. Brennan, J. (2016). The ethics and rationality of voting. In E. N. Zalta (Ed.), The Stanford Encyclopedia of Philosophy (Winter 2016 ed.). Metaphysics Research Lab, Stanford University. Christiano, T. (2018). Democracy. In E. N. Zalta (Ed.), The Stanford Encyclopedia of Philosophy (Fall 2018 ed.). Metaphysics Research Lab, Stanford University. Dowding, K. (2005). Is it rational to vote? Five types of answer and a suggestion. The British Journal of Politics and International Relations 7(3), 442â&#x20AC;&#x201C;459. Downs, A. (1957). An Economic Theory of Democracy. New York: Harper and Row. Elster, J. (2007). Explaining Social Behavior: More Nuts and Bolts for the Social Sciences. Cambridge: Cambridge. Frazier, R. (1998). Duty. In Routledge Encyclopedia of Philosophy. Jeske, D. (2019). Special obligations. In E. N. Zalta (Ed.), The Stanford Encyclopedia of Philosophy (Fall 2019 ed.). Metaphysics Research Lab, Stanford University. Lomasky, L. E. and G. Brennan (2000). Is there a duty to vote? Social Philosophy and Policy 17(1), 62. Mackie, G. (2010). Why itâ&#x20AC;&#x2122;s rational to vote. Unpublished ms. Maskivker, J. (2019). Why there is a moral duty to vote. In OUPBlog.

Democratic Duty to Vote

Parfit, D. (1984). Reasons and Persons. Oxford: Oxford. Tuck, R. (2008). Free Riding. Cambridge, MA: Harvard.

What is MacKinnon’s Argument Against Pornography? rebecca clark

HILOSOPHICAL critiques of pornography are like ‘arguing with an orgasm’, as Catharine MacKinnon bluntly put it (MacKinnon 1993: 17). Yet despite this seemingly insurmountable challenge, MacKinnon launched a scathing attack on pornography in her seminal text Only Words (MacKinnon 1993). In this essay, I will outline MacKinnon’s conception of pornography before laying out her three main critiques of it: first, that it is an instance of rape of female performers; second, that it constitutes the subordination of women; and third, that it silences women. I will argue that much of the subsequent academic literature, in fetishizing Austin, has obscured MacKinnon’s crucial contribution to the pornography debate— that whether pornography is treated as mere speech is not a theoretical matter to be settled in the philosophy of language, but an ideological question currently decided by patriarchal institutions.

1 Preliminaries What is pornography? MacKinnon defines it as ‘graphic sexually explicit materials that subordinate women through pictures or words’ (MacKinnon 1993: 22); this can be contrasted with an ostensibly more neutral definition of pornography, roughly understood as sexually explicit material with the primary intent to arouse (Finlayson 2015: 137). Mikkola rejects MacKinnon’s definition of pornography in favour of such a ‘non-evaluative definition’ on the basis that feminist theorists, by adopting an ordinary understanding of the term, can better shape public policies around pornography (Mikkola 2019: 17). However, there are two key aspects of Mikkola’s claim which I wish to challenge. Firstly, it is an evaluative judgement to keep open the question of

British Undergraduate Philosophy Review 1 (2020)

MacKinnon on Pornography

whether or not pornography is wrong—for instance, one might think that pornography is just so obviously wrong that it is appropriate to incorporate a negative stance on pornography in its very definition—and hence no definition of pornography can claim supposed ‘neutrality’ (Finlayson 2015: 139). Secondly, Mikkola’s claim that a non-pejorative definition of pornography is methodologically superior in the sense of having a bigger impact on real-world debate is fundamentally disputed by MacKinnon. To put it crudely, MacKinnon’s ‘methodology’ is to motivate a Gestalt shift whereby her readers experience a sudden change in their perception of pornography and its relation to the social world. Her use of rhetoric is not methodologically unsound but rather vital in achieving this. It is worth noting a few further things about MacKinnon’s approach. Firstly, she rejects a moralistic account of what is wrong with pornography such as that it promotes obscenity, as adopted by many conservatives (MacKinnon 1983). Instead, she argues on political grounds that pornography is a form of sex discrimination which undermines women’s civil rights. Secondly, empirical data does not feature centrally in MacKinnon’s writing since she considers pornography’s influence on society to be pervasive and fundamental in structuring the collective social imaginary, and hence the full extent of its effects cannot be neatly captured in an empirical study. In its place, MacKinnon deploys shocking and incisive interpretations of what happens to women in pornography, since it is reasonable to expect what pornography is like to influence what it can do.

2 Rape, subordination, and silencing MacKinnon’s first key argument against pornography is that it necessarily involves the rape of the women used in its production. According to her radical redefinition of rape, any sex which involves the exploitation of a power differential is rape (MacKinnon 2016). She claims that ‘all pornography is made under conditions of inequality based on sex’(MacKinnon 1993: 20), and that without these background conditions of inequality, no woman would freely choose to do pornography. Hence in MacKinnon’s view, a woman who believes herself to have voluntarily agreed to have sex on camera in exchange for payment has in fact been raped since she only agreed to do so due to her disadvantaged position. This redefinition would make for an unsatisfactory legal account of rape since it is impossible to determine what people would want in a radically different world, given that we are partly constituted by our desires; however, I am intuitively sympathetic to it as an ethical account. It is worth noting that this is

Rebecca Clark

the most optimistic picture of how women get involved in the pornography industry, with many physically coerced into it and thus raped under the conventional use of the term.1 MacKinnon’s second main contention is that pornography constitutes subordination. Indeed, the claim that pornography is a form of subordination, as opposed to just depicting or perpetuating it, is embedded in MacKinnon’s very understanding of the term. In 1993, Rae Langton attempted to demonstrate its philosophical plausibility in her pivotal article ‘Speech Acts and Unspeakable Acts’ by drawing on Austin’s speech act theory. Roughly, a speech act is an act you can perform simply by saying that you are doing so (although this is not always a necessary condition for performing the speech act) (Green 2020). Consider the parent who wants to encourage their child to do their homework. If the parent says, ‘I am persuading you that you need to do your homework’, this evidently does not constitute persuading the child of this. However, if the parent says, ‘I am warning you: if you don’t do your homework, you’re not allowed to watch television this evening’, then in saying this, the parent has performed a speech act of warning. Following Austin, Langton identifies three dimensions of a speech act: its locutionary, perlocutionary, and illocutionary acts. A locutionary act is ‘to utter a sentence that has a particular meaning, as traditionally conceived’; a perlocutionary act is ‘the action performed by saying something’; an illocutionary act is ‘the action performed simply in saying something’ (Langton 1993: 295, 300, 301). For instance, suppose I were to say ‘I do’ as the bride in a marriage ceremony when motioned by the priest. The locutionary act is the literal meaning of ‘I do’ (referring to myself consenting); the perlocutionary act is the effects of saying ‘I do’ (for instance, my mother starts weeping); the illocutionary act is the act performed in saying ‘I do’, i.e. that I am thereby married. A speech act has illocutionary force when it satisfies certain so-called ‘felicity conditions’ which are set by conventions—for instance, in the case of marriage, until 2013 there was a felicity condition in England that the couple must be heterosexual for the words ‘I do’ to have illocutionary force. Langton deploys speech act theory to show that it is philosophically coherent to claim that pornography is an illocutionary act of subordination (although she herself does not necessarily commit to the view that it is such an act). Her argument can be roughly formulated as follows:2 1 Although, as sex workers Mac and Smith point out, the standard narrative of sex-trafficking as in-

volving the kidnapping of women is itself far from accurate (Smith and Mac 2018: ch. 3) 2 Note that my use of [...] denotes a slightly distinct argument which holds that a speech act only

MacKinnon on Pornography

P1: certain types of speech are speech acts P2: pornography is a speech act P3: a speech act is an illocutionary act of subordination if it meets [at least some of] the felicity conditions for subordination P4: the felicity conditions for subordination are: (1) the speech act does one or more of the following: (a) unfairly ranks members of a social group as inferior (b) legitimates discriminatory behaviour towards them (c) unjustly deprives them of some important powers (2) the speaker’s intentions are satisfied (3) the speaker achieves uptake (i.e. the hearer recognizes the intended illocution being performed) (4) the speaker is authoritative in the relevant domain P4: pornography satisfies [at least some of] the felicity conditions for the subordination of women C: therefore, pornography subordinates women In this argument, pornography is taken to be an illocutionary act of subordination since it ranks women as sex objects and legitimates discriminatory behaviour in the form of systematic sexual violence against women. However, Langton herself acknowledges that pornography falls short of the illocutionary paradigm since it is not clear that (2), (3), or (4) are satisfied—pornographers do not necessarily intend to subordinate women, pornographic speech may fail to achieve ‘uptake’ if consumers do not recognize it as subordinating women, and pornography may not to be authoritative in the sexual domain (although this is the most plausible claim out of the three). Nevertheless, Langton suggests that MacKinnon’s subordination claim could be saved—perhaps pornography is close enough to the illocutionary paradigm to count as an instance of subordination, or it is the best explanation for pornography’s perlocutionary effects. needs to satisfy some felicity conditions for subordination for it to be an illocutionary act of subordination (‘close enough’ argument).

Rebecca Clark

Before addressing this debate, I wish to consider MacKinnon’s final and closely related argument that pornography constitutes the silencing of women. As she succinctly puts it, pornography has ‘deprived women of speech, especially speech against sexual abuse’ (MacKinnon 1993: 9). As before, Langton steps into the picture to elucidate MacKinnon’s silencing claim. According to Langton, there are three types of silencing: failure to perform a locutionary act, perlocutionary frustration, and illocutionary disablement (Langton 1993: 315). I will focus on the third type of silencing since this is the most controversial of MacKinnon’s claims, although she considers pornography to cause all three. Illocutionary disablement is when one’s speech fails to perform the very action that one intends (Langton 1993: 322). Women may be unable to perform certain illocutionary acts such as refusal and protest, particularly in sexual contexts, as ‘the felicity conditions for women’s speech acts are set by the speech acts of pornography’ (Langton 1993: 324). For instance, Langton asks us to consider the case in which pornography instils rape myths in men to such a degree that when a woman says ‘No!’ before sex, that is honestly interpreted as meaning yes by a man. In such a case, the woman has failed to secure the uptake required for her speech to constitute the illocutionary act of refusal, and hence she is silenced—and raped (Langton 1993: 320–321). Whilst this rape scenario has received much attention in the literature, Finlayson stresses that it is merely a model used to elucidate the way in which our social context determines what speech acts are and are not possible (Finlayson 2014: 781). Thus it is not a decisive rebuttal of Langton’s argument to point out that in the vast majority of rapes, the perpetrator is aware that the victim does not in fact consent to sex, since this example is being employed pedagogically to highlight the way in which a woman’s words can be twisted and misinterpreted. Consider, for instance, a woman subject to unwanted advances from a man who fails to pick up on her stiff body language and curt responses to his questions. It is not the case that the woman cannot perform the speech act of refusal; rather, to do so would require her to escalate the situation, which carries associated risks such as the man reacting aggressively. This kind of situation is a function of a patriarchal system in which women’s speech is consistently undermined, to which pornography can be said to contribute.

3 Corn-dealers as deal-breakers An opinion that corn-dealers are starvers of the poor, or that private property is robbery, ought to be unmolested when simply circulated through

MacKinnon on Pornography

the press, but may justly incur punishment when delivered orally to an excited mob assembled before the house of a corn-dealer. (Mill 2003: 131) Many analytic feminist philosophers have contested the claim that pornography is an illocutionary act of subordination or silencing—as opposed to ‘merely’ a perlocutionary one—on various grounds (Mikkola 2019: ch. 2–3). However, I concur with Finlayson that couching MacKinnon’s subordination and silencing arguments against pornography in Austinian terms, whilst a useful contribution to speech act theory, fundamentally misrepresents her core argument. For MacKinnon, there is no deep ontological distinction between illocutionary and perlocutionary speech acts at all. As she herself acknowledges, ‘both illocution and perlocution are causal theories, the former more immediately and with fewer intervening contingencies than the latter’ (Maitra and McGowan 2012: foreword). Given this, what are we to make our MacKinnon’s claims? Finlayson contends that ‘by identifying porn with subordination (rather than following the more usual practice of identifying it as a cause of subordination), MacKinnon is making a claim about the intimacy, immediacy and systematicity of the relationship she sees between pornography, on the one hand, and violence and discrimination against women, on the other’ (Finlayson 2014: 784, emphasis added). In projecting the idea that women are sex objects who derive pleasure from cruelty, dehumanization and rape, pornography constructs ‘the social reality of what a woman is and can be in terms of what can be done to her, and what a man is in terms of doing it’ (MacKinnon 1993: 25) and shapes the standard experience of sex itself. In this way, pornography is closely interconnected with the subordination and silencing of women. Ultimately, MacKinnon’s second and more fundamental Gestalt shift is in enabling us to see that whether speech acts are regarded as perlocutionary or illocutionary speech is itself a normative question. Consider Mill’s famous example of corndealers: supposing that in both contexts the claim that corn-dealers are starvers of the poor could conceivably lead to them being violently assaulted, what is the distinction to be made between these two speech acts?3 MacKinnon exposes Mill’s distinction as a choice of whether to consider the effects of a speech act to be part of the act or merely a contingent causal consequence of it. As MacKinnon notes, we legally treat many forms of speech as if it were obviously not just speech. For instance, if a woman were to shout ‘Shoot!’ to someone pointing a 3 As

Simpson and Srinivasan note, this distinction ‘starts to look tenuous on inspection’ (Simpson and Srinivasan 2018: 191).

Rebecca Clark

gun at a child, she would be recognized by the state as an accomplice in the crime ‘without so much as a whimper from the First Amendment’ (MacKinnon 1993: 12). Yet despite certain types of pornography effectively calling for ‘Rape!’, any mention of it subordinating women inevitably descends into a fierce dispute over free speech. More than anything, this reveals how the notion of ‘free speech’ is wielded as an ideological weapon in pornography debates to protect men’s superior social standing. Building on this, Bauer suggests that the question of whether a speech act is perlocutionary or illocutionary is in fact an ethical one of where to assign responsibility (Srinivasan 2018: 1413). Hence the real question is not whether pornographic speech harms women, but whether we want to hold it accountable.

4 Conclusion With the advent of widespread internet pornography, MacKinnon’s arguments are more relevant than ever before. Her withering three-fold critique of pornography centring on the rape, subordination and silencing of women brings to the foreground its role in socially constructing a world in which women are perceived as sexual objects, violence against women is sexualized, and women’s refusals are silenced. The past few decades have seen MacKinnon’s views forced into an Austinian straitjacket, masking her finest insight that the reified distinction between illocutionary and perlocutionary is largely illusionary. The question we must now ask is who should be held responsible for the sexual subordination and silencing of women, in which pornography plays a key part. Acknowledgements I am grateful to Professor Amia Srinivasan for valuable conversations on the issues with which this article is concerned. Balliol College, Oxford, England

References Finlayson, L. (2014). How to screw things with words. Hypatia 29(4), 774–789. Finlayson, L. (2015). An Introduction to Feminism. Cambridge. Green, M. (2020). Speech acts. In E. N. Zalta (Ed.), The Stanford Encyclopedia of Philosophy (Winter 2020 ed.). Metaphysics Research Lab, Stanford University.

MacKinnon on Pornography

Langton, R. (1993). Speech acts and unspeakable acts. Philosophy and Public Affairs 22(4), 293–330. MacKinnon, C. A. (1983). Not a moral issue. Yale Law & Policy Review 2, 321. MacKinnon, C. A. (1993). Only Words. Harvard. MacKinnon, C. A. (2016). Rape redefined. Harvard Law & Policy Review 10, 431. Maitra, I. and M. K. McGowan (2012). Speech and Harm: Controversies Over Free Speech. Oxford. Mikkola, M. (2019). Pornography: A Philosophical Introduction. Oxford. Mill, J. S. (2003). Utilitarianism. In M. Warnock (Ed.), Utilitarianism and on Liberty, Chapter 4, pp. 181–235. Wiley. Simpson, R. M. and A. Srinivasan (2018). No platforming. In J. Lackey (Ed.), Academic Freedom, pp. 186–209. Smith, M. and J. Mac (2018). Revolting Prostitutes: The Fight for Sex Workers’ Rights. Verso. Srinivasan, A. (2018). How to do things with philosophy. European Journal of Philosophy 26(4), 1410–1416.

Phenomenal Overﬂow: What an Auditory Analogue of Sperling Does Show alexander gruen

1 Introduction HE nature of our conscious experience goes to the very heart of the philosophy of cognitive science. Recently, much of this debate has centred around the distinction between access and phenomenal consciousness, introduced by Block (1995). In this paper, I will present a new argument showing that we may have phenomenal consciousness without access consciousness. While I argue that this ‘phenomenal overflow’ is not necessarily as extensive as Block imagines, this still leaves room to believe that his basic schema is the best explanation for the findings from a number of partial report experiments such as Sperling’s (Sperling 1960). We will begin by outlining Block’s distinction between phenomenal and access consciousness and explain his argument for phenomenal consciousness without access in Sperling’s experimental paradigm. I will then address the influential objection by Phillips (2011), who argues that postdiction can account for these findings. Phillips’ work has provoked much further research, with researchers aiming to find postdictive effects which may be able to explain Sperling’s findings (for example, Sergent et al. 2013). While there is a live debate as to whether these attempts have been successful, I will not address this here. Instead, I will present Darwin et al. (1972) auditory analogue of Sperling’s experiment, which involves much longer time delays than Sperling’s original experiment. This has not been used to argue for phenomenal overflow before. However, due to the much longer timescales involved, it is significantly harder to explain these findings using postdiction. As such, these findings provide a stronger basis for phenomenal overflow than Sperling’s results. However, while this auditory effect lasts for much longer, it is significantly reduced in scale compared to its visual analogue. This supports the ultimate conclusion that we have overflow, but perhaps not as much as Block thinks.

British Undergraduate Philosophy Review 1 (2020)

Phenomenal Overﬂow

2 Phenomenal and access consciousness Ned Block introduced two different types of consciousness in his seminal paper ‘On a Confusion about the Function of Consciousness’ (Block 1995). He argued that many of the philosophical disagreements about the nature of consciousness could be attributed to a failure to distinguish between these two distinct concepts. The first type of consciousness that Block introduces is phenomenal consciousness (P-consciousness). He believes that we cannot define this in any non-circular way. As he notes though, the history of reductive definitions in philosophy is such that we should not necessarily expect such a definition for anything, let alone a concept as fraught as consciousness. Accordingly, the best we can do is point to phenomenal consciousness by way of synonyms and examples (Goldman 1993). Roughly speaking, a state is phenomenally conscious if there is ‘something it is like’ to be in that state.1 On this understanding, phenomenal consciousness is experience. There are also myriad examples of phenomenally conscious mental states. For example, we have phenomenally conscious mental states whenever we see, hear, smell or taste things as well as when we are in pain. Phenomenally conscious properties include the experiential properties of sensations, feelings, perceptions, thoughts and emotions. The second type of consciousness that Block outlines is access consciousness (Aconsciousness). A mental state is access conscious if it is available for use in reasoning and rationally guiding action. Block talks about both states and their contents as access conscious. We say that people are access conscious (of something) in light of having access conscious mental states. For example, when a subject sees a red light, they are access conscious of it if they can use this perception to rationally guide their action. If they are approaching the light in a car, a natural rational response would be to slow down and stop, rather than continue on through the intersection (although there may, of course, be some reckless drivers who would continue on despite their access to the knowledge that the light is red). There has been a great deal of debate in contemporary philosophy about the precise relationship between these two different formulations of consciousness. I will call the claim that we can have phenomenal consciousness without associated access consciousness the thesis of ‘phenomenal overflow’. This thesis implies that the contents 1 This idea is heavily influenced by Nagel, who famously wrote, ‘...fundamentally an organism has conscious mental states if and only if there is something that it is like to be that organism - something it is like for the organism.’ (Nagel 1974: 436)

Alexander Gruen

of phenomenal experience are in some sense richer than, or overflow, what we can access. I will argue that we perhaps have a reduced level of overflow compared to what Block imagines but will ultimately defend an overflow thesis.

3 Block’s argument for phenomenal overﬂow Block gives a number of supposed examples of phenomenal overflow. The most widely discussed and successful of his arguments relies on the experimental results of Sperling (Sperling 1960). In Sperling’s most famous experiment, subjects were shown arrays of 12 letters for 50 milliseconds, arranged into three rows of four letters (see fig. 1).

Figure 1: A Sperling-type array When asked immediately to report all the letters that they saw (the whole report condition), subjects could report an average of 4.3 letters out of the 12 they were shown (an average of just over 1.4 letters per row). However, in the partial report condition, a tone was played to cue the subjects to report only one of the rows (a high tone corresponding to the top row, a medium tone to the middle row and a low tone to the bottom row). When this cue was played immediately after the display was offset, subjects could recall about 3 of 4 letters in a given row—significantly more than the (on average) 1.4 letters they could recall from each row in the whole report condition. We will call this the partial report advantage. However, in some trials, the cue was delayed by varying amounts of up to a second. Sperling found that there was still a robust partial report advantage with delays of at least 300 milliseconds, with subjects being able to report (on average) 2 letters from a given row in this particular partial report condition. Block has used these results to mount an argument for the presence of phenomenal experience without associated access. The argument appeals to partial report advantage as well as subjects’ reports. Subjects in these experiments often report that they could see all or almost all of the letters in the array (Baars 1988; Sperling 1960). How are we to square this with the fact that subjects are only able to correctly identify about half of them?

Phenomenal Overﬂow

Block claims that the answer lies in acknowledging a separation between A and P-consciousness. When subjects are cued to report just a particular row of the array, they can report all or nearly all the letters in this row. This suggests phenomenal experience as of these letters. This is the case even when they are cued shortly after the visual stimulus is offset. Given the cue would not be expected to retrospectively affect a subject’s experience, we can expect that the subject was phenomenally conscious of a similar number of letters in the other rows. As a result, we come to the conclusion that the subject is phenomenally conscious of all or nearly all the letters in the array. Importantly, this phenomenology is specific enough to allow a subject to accurately report any given one of these letters (if they are cued appropriately). On this view, the temporary visual store posited by Sperling is phenomenal in character and corresponds with phenomenal experience for the subject. However, Block goes on to argue that at any given time subjects are not access conscious of each letter in a similar way. Subjects are not able to report more than about four letters in either the whole or the partial report condition, so the unreported letters are not available to the subject for use in rational decision making. There is no time when a subject has access to all or nearly all the letters jointly. Subjects are of course, access conscious of the fact that there are 12 letter-like forms before them, but this A-consciousness is not specific enough to allow subjects to identify all the letters in the array at any given point in time. While this is a very specific example of P-consciousness without A-consciousness, Block believes that this it is not uncommon for these two types of consciousness to dissociate. To back up this position, he points to the magnitude of the partial report advantage. For Block, we have phenomenal experience of all or nearly all of the array and can only remember some details about it. Similarly, every time we look around a room, we are likely to experience much more than we can report. So, the question of whether we have P-consciousness without A-consciousness is really about the nature of our everyday experience, rather than solely about certain subjects in a laboratory in 1960.

4 What Sperling doesn’t show One of the most influential responses to Block was put forward by Phillips (2011), who gave an alternate explanation for these experimental findings involving perceptual postdiction, aiming to undermine Block’s argument as an inference to the best explanation. Phillips specifically points out two assumptions that he claims are minimally

Alexander Gruen

necessary to ground Block’s argument. The first assumption is that subjects’ reports do indeed connect to conscious experience. He is willing to grant this assumption, as subjective reports are ‘paradigmatic criteria for awareness’ (Phillips 2011: 385). To deny that such reports are strong presumptive evidence of subjective experience would be ‘to lose experimental purchase on conscious experience’ (Phillips 2011: 385). The second assumption that Block’s argument relies upon is that it is legitimate to sum partial reports to ascertain how many letters subjects were P-conscious of in the array. This assumption relies on the following counterfactual, given by Phillips: (CF) Any aspect of experience present in a partial report condition would have been present even if some other partial report had been cued. (Phillips 2011: 385) If we have this counterfactual, we can argue that it is legitimate to sum partial reports as follows: 1. A subject can usually report about 3 out of 4 letters in a cued row, which is indicative of phenomenal consciousness of these 3 letters (by our first assumption). 2. If a different row had been cued, they would similarly have been able to report about 3 of the 4 letters in it, indicating conscious experience of about 3 letters in each uncued row. 3. Conscious experience is not altered by the choice of cue (by CF). 4. So, subjects consciously experience (about) 3 of the letters in each of the rows that are not cued, as well as the 3 they experience in the one that is. C: So, subjects are phenomenally conscious of about 9 letters in the array. This argument was implicitly used above by Block, but this formulation makes clear the role that the two assumptions play in justifying the conclusion. These two assumptions are the only controversial parts of this argument: the rest of the argument is simply outlining experimental findings. So, if we accept both these assumptions, we should accept the conclusion. However, Phillips wishes to push back against the assumption of CF. Specifically he highlights that CF is based on an ‘independence assumption’ - the assumption that

Phenomenal Overﬂow

the conscious experience of the array is not affected by (and is, in this sense, independent of) the cue tone. This was justified above by appealing to the fact that the cue tone is played after the array is offset. However, this does not, in fact, guarantee the independence of the cue tone and our perception. Phillips appeals to a phenomenon known as perceptual postdiction to demonstrate this. This phenomenon involves a stimulus which is presented later appearing to causally affect the perception of another stimulus which was presented earlier. We will generally call the stimulus which is presented later the ‘moderator’ or ‘masking’ stimulus. The ‘target stimulus’ is the stimulus that the ‘moderator’ or ‘masking’ stimulus may appear to alter the perception of (the earlier stimulus in this definition). Phillips argues that the experimental evidence of postdictive effects provides good reason to doubt the independence assumption and thus CF. On this reading, the cue tone influences our perception of the array, and thus it is not legitimate to sum partial reports in order to establish phenomenal consciousness of all the letters jointly. While he accepts that subjects have phenomenal experience of all the letters they report, he argues that because their experience of the array may be cue-sensitive, we cannot draw conclusions about phenomenal experience of letters in the uncued rows from performance in the partial report condition. He does not argue directly that subjects’ experience of the array is, in fact, cue-sensitive, but he does aim to show that it is, at the least, not an unlikely explanation of Sperling’s findings. In order to motivate the hypothesis that the perception of the array is cue-sensitive, Phillips presents a number of postdictive phenomena. He argues that these postdictive effects occur in contexts which are similar in important ways to Sperling’s experimental paradigm, and thus postdiction may explain the partial report advantage. By providing an alternative explanation of Sperling’s findings, Phillips aims to show that we can explain them without appealing to the unnecessary assumption that subjects’ experience of the array is not cue-sensitive. Consequently, Phillips claims to undermine Block’s argument that Sperling’s experiments provide a clear case of phenomenal consciousness without access. There has been a great deal of debate as to how successful Phillips’ argument is. Most would agree that at the least Phillips has raised a salient point that requires further investigation. There isn’t a consensus as to whether Sperling’s experiments demonstrate phenomenal overflow. I will not weigh into this debate here, although I do note that it is very hard to conclusively rule out postdiction as an explanation for Sperling’s experiments. As Phillips points out, there are a number of postdictive

Alexander Gruen

phenomena which occur in paradigms similar to Sperling’s (such as sound-induced visual bounce (Sekuler et al. 1997) and the flash-lag illusion (?? Eag)). As such, there is often a certain dissatisfaction for defenders of an overflow thesis in being unable to conclusively rule out postdiction as an explanation for Sperling’s results.

5 What an auditory equivalent of Sperling does show As a result, I will take a different tack, presenting an argument for phenomenal overflow based on the findings of Darwin et al. (1972), who conducted an auditory analogue of Sperling’s experiment. They demonstrated the existence of a brief auditory store (commonly termed ‘echoic memory’ (Neisser 1967)) analogous to the brief visual store (iconic memory) demonstrated by Sperling. However, echoic memory was found to persist for significantly longer than iconic memory, making postdiction an implausible explanation of these findings, and undermining Phillips-style objections to our overflow conclusion. In this experiment, subjects were played auditory stimuli through headphones with partial reports elicited by a visual cue. They were simultaneously presented with a different list of three consecutive items from each of three ‘spatial’ locations (left ear, middle, right ear). The stimulus considered to be in the ‘middle’ was played in both the left and the right ears. Items in each list were selected from the letters BFJLMQRUY and the numbers 1 to 10 (excluding seven as it is disyllabic in contrast with all the other letters and numbers selected). Each list was at a rate of three items per second. These lists were then combined to create a stimulus for each trial consisting of one list which was played through both headphones, one list played only through the left headphone and another list played only through the right headphone. Trials were split by reporting condition, with blocks being composed either of full report trials or partial report trials. In the whole report condition, subjects were asked to report as many items as they could from the nine that they were presented with (three lists of three items) and were allowed to respond in their own time. In the partial report condition, subjects were cued either zero, 1, 2 or 4 seconds after the end of the stimulus to report only one of the three lists played to them by its spatial location. A visual cue was used: subjects were cued using a blank bar on a slider on screen in front of them similar to those shown in Figure 2 below. If the bar was on the left, this would cue report from the left list, and similarly if the bar was in the middle or on the right this would cue report from these spatial locations. We can see, from Figure 3 that in the whole report condition, subjects could report

Phenomenal OverďŹ&#x201A;ow

32 Left cue

Middle cue

Right cue

Figure 2: The type of visual cue used to elicit partial reports in Darwin et al. 1972. The left cue above corresponds to the visual cue for the left list, and similarly for the middle and right cues. roughly 4.2 of the 9 items that they were presented with, corresponding to approximately 1.4 items from each 3-item list. In the 1 and 2 second delayed partial report conditions, subjects reported between 1.5 and 1.6 items from each 3-item list. Now, the first thing that is noticeable about these results is how small this partial report advantage is. In the visual case, this partial report advantage was significantly larger, with subjects being able to identify roughly 1.4 items from each row in Sperlingâ&#x20AC;&#x2122;s array in a full report condition, compared to slightly over 3 items from each row in some partial report conditions. The authors give some possible explanations for this significantly reduced effect relative to the visual case: The most plausible reason [for this reduced effect] is that the subjects found it difficult to separate the three different input channels. Many of the errors were intrusion errors from different channels and a number of the subjects said that they had difficulty distinguishing the middle channel. This suggests that subjects would have performed rather poorly if they had been precued to select one channel. By contrast, in the visual modality performance is very good with precueing by location (Eriksen & Johnson, 1969). It may be possible to improve the separation of the three channels by using a different voice on each channel. (Darwin et al. 1972: 258) So, we might be able to produce a greater effect by changing our experimental paradigm slightly. Also, as the authors suggest, in this experiment the different channels (left, right and middle) are very hard to distinguish. We can see this in the immediate (zero second delay) partial report condition, where subjects are still only able to

Alexander Gruen

Figure 3: Results from Darwin et al. (1972), with the black line representing the average partial report at the various cue delays multiplied by 3 (corresponding to the summation of partial reports over the 3 different cue conditions). Adapted from Darwin et al. 1972. report a little over 1.6 items from each list. Further, while we donâ&#x20AC;&#x2122;t have a pre-cueing condition, the authors suggest that they would expect performance to still be quite poor. Correspondingly, even a small partial report advantage (of 0.2 letters per row) is impressive given subjectsâ&#x20AC;&#x2122; poor performance on the immediate cue condition and expected poor performance in a pre-cueing condition. Even this small partial report advantage represents a significant jump (in percentage terms) from our baseline performance (whole report) to the best performance we could record for the task if we trained all our attention on the appropriate list (in a pre-cueing condition). As the difference between whole and partial report is relatively small, we might doubt whether there is a real effect at play here. However, according to Darwin et al., we have a highly statistically significant effect for the 0, 1 and 2 second cue delays (all with a p-value of less than 0.001). These p-values indicate a less than 1 in 1000 chance of witnessing these levels of partial report advantage as a result of random chance if there was no widespread partial report advantage. Further, all subjects exhibited par-

Phenomenal Overﬂow

tial report advantage with each of these cue delays, further reinforcing the legitimacy of this effect. So, we can conclude that although this effect may seem small, we have a legitimate partial report advantage on timescales up to at least 2 seconds. The 4 second cue delay does not give a statistically significant partial report advantage, even though average performance was slightly better in the 4 second delay cue condition than in the whole report condition. The existence of this brief auditory store allows us to adapt Block’s argument for overflow: 1. A subject can usually report about 1.57 out of 3 items in a cued list, which is indicative of phenomenal consciousness of these items. 2. If a different list had been cued, they would similarly have been able to report about 1.57 of the 3 items in it, indicating conscious experience of about 1.57 items in each uncued list. 3. Subjects’ conscious experience of the auditory stimulus is not altered by the choice of visual cue. 4. So, subjects consciously experience (about) 1.57 of the items in each of the lists that are not cued, as well as the 1.57 that they experience in the one that is. C: So, subjects are phenomenally conscious of about 4.7 items from the stimulus but can only access 4.2 of these letters at any given time. So, the argument for overflow based on the auditory analogue of Sperling is very similar to Block’s original argument. The essential differences between the experiments that these arguments are based on can be summarized as follows: 1. Sperling’s experiment involves the reporting of a visual stimulus (with an auditory cue), which differs from the auditory report studied by Darwin et al.. 2. Subjects in the auditory analogue of Sperling exhibited statistically significant partial report advantage over a much longer timescale than subjects in Sperling’s experiment. 3. The partial report advantage in the auditory analogue of Sperling experiment is significantly smaller than in Sperling’s experiment.

Alexander Gruen

Now, as outlined earlier, Phillips points out that an argument for phenomenal overflow based on partial report superiority requires at least two key assumptions. It is worth recalling what these are. The first assumption is that subjects’ reports in an experiment do indeed relate to conscious experience. The second is that it is legitimate to sum partial reports to ascertain how many letters subjects were conscious of in the sensory stimulus. This assumption relies on the counterfactual given above as CF. We need to justify both assumptions for this altered experimental paradigm if we wish to make an argument for phenomenal overflow. In Sperling’s experiments, Phillips considers it legitimate to assume that if a subject reports seeing the presence of a feature, this is strong presumptive evidence that the subject has had conscious experience as of that feature. This is because he considers such reports as ‘paradigmatic criteria for awareness’. As such, in order to maintain an experimental handle on consciousness, he believes that we should grant this assumption. This argument also seems to apply to the auditory analogue of Sperling. When reporting on what letters or numbers they heard in these experiments, subjects were selecting one item from a list of 18 possible items (9 letters and 9 numbers). This task requires a high level of specificity, much like reporting which letter (out of the 26 in the alphabet) was seen in Sperling’s experiment. The level of discrimination required in both paradigms is significant. It is certainly not the case that subjects were participating in a forced choice test, which can often overestimate conscious experience (Phillips 2011; Voss et al. 2008). So, if we accept, as Phillips does, that reports in Sperling’s experiment are paradigmatic criteria of visual awareness, it seems that in much the same way, reports in these experiments should be viewed as paradigmatic criteria of auditory awareness. So, our first assumption is justifiable. The second assumption is the more controversial of our two assumptions. Now, the first worry we need to address here is the size of our partial report advantage. As we pointed out above, the size of the auditory partial report advantage is much smaller than in the visual case. Given the reduced magnitude of this effect, there might be some doubt about whether it can really ground an argument for phenomenal overflow. Specifically, the summing of partial reports at 1-2 second cue delays gives subjects awareness of 4.7 items in the stimulus, where they can usually report about 4.2 items in the whole report condition. However, this effect is statistically significant with a p-value of less than 0.001 as outlined above, so needs explaining. It can’t be dismissed as evidence just because the magnitude of the effect is small. Any competing theory of conscious experience needs to be able to explain this effect, despite its size.

Phenomenal Overﬂow

Now, perhaps Phillips might argue that such a small effect could be due to some sort of subconscious processing of the auditory stimulus rather than phenomenal experience. If we accept our first assumption, such an explanation of the results is really an account of perceptual postdiction. It involves a change in the perceptual content of our experience based upon which cue we choose. So, if we can show that postdiction is again a very unlikely explanation for this phenomenon, we will have a strong argument for some form of phenomenal overflow, even though the magnitude of this overflow is much smaller than is claimed in the visual case. It is not the magnitude of the effect that is important as it pertains to our overflow argument; the important thing is that there is a legitimate partial report advantage over a timescale of at least 2 seconds. So, on what sort of timescale can we have postdiction? Phillips provides numerous examples of postdictive phenomena occurring when the moderator stimulus is presented within a couple of hundred milliseconds of the target stimulus, but no examples on timescales longer than about half a second. Certainly, he provides no effects which operate over multiple seconds. Further, it is widely accepted that there is a temporal limit to postdictive perceptual effects, often assumed to be in the realm of a few hundred milliseconds. In the words of Shimojo, These phenomena will clearly suggest that there is a limited temporal time range (on an order of 100–200 ms) within which the processing of a stimulus presented later can affect the percept of another stimulus presented earlier. (Shimojo 2014: 196) So, even if we are generous with the timescales on which postdiction may be present, it doesn’t get us near two seconds. It is also worth noting that Phillips himself acknowledges that postdiction is an unlikely explanation for any kind of partial report advantage with cue delay of multiple seconds. When talking about an experiment aiming to show fragile visual short-term memory (VSTM), he acknowledged that, The cue delays involved in these experiments are significantly longer than in Sperling’s partial report task: at least 1s, and with sufficient training perhaps up to 4s. Given this, it is far less plausible to account for the findings in terms of perceptual postdiction. (Phillips 2011: 403) So, it seems unlikely that Phillips would claim that we have postdiction in the auditory analogue of Sperling, given that partial report superiority in this paradigm

Alexander Gruen

persists until at least 2 seconds. The partial report superiority persists on a timescale which is an order of magnitude larger than the 200 milliseconds or so at which most postdictive effects are significant. As such, we could allow that these postdictive effects last considerably longer than 200 milliseconds and still be well short of the two seconds required to explain the findings of Darwin et al. (1972). We can, of course, never prove conclusively that postdiction is not as play here, but, it is evident that postdiction is a very implausible explanation of effects such as this one, which last well beyond one second. As such, we have presented a strong argument for phenomenal consciousness in the absence of access, as we have justified both assumptions needed to move from partial report superiority to phenomenal overflow.

6 Conclusion There has been a great deal of debate over the detail of Sperling’s experiments and whether they can ground an argument for overflow that is resistant to Phillips’ objection. However, this paper takes the debate beyond this point by venturing beyond visual, to auditory stimuli. While it remained feasible to defend postdiction as an explanation for Sperling’s results in the visual case, we’ve argued that the sheer length of time for which the partial report advantage is significant makes this implausible in the auditory case. The partial report advantage in this experiment is much less significant, and thus grounds an argument for less overflow than originally envisioned by Block. Thus, by using evidence from a quite different perceptual domain, we corroborate recent experimental findings which suggest that even in the visual case, we may have less postdiction than Block imagines (Sergent et al. 2013). This paints a picture on which we may have rich conscious experience that surpasses report, but only slightly. If these arguments are right, we do have phenomenal overflow, but only at the fringes of our experience. The fact remains however that any appreciable overflow has significant philosophical import. We have not explored the consequences of our phenomenal overflow thesis, as the focus of this paper was on demonstrating the presence of overflow. While we have argued for a reduced level of overflow compared to what Block envisions, our view shares many similar metaphysical consequences with Block’s. This has significant implications for our understanding of our everyday experience and even our notion of subjectivity and the relationship between an experience and its subject. Thus, by paring back our view of overflow, we have preserved some of the key conclusions reached in Block’s initial work.

Phenomenal Overﬂow

Acknowledgements I would like to thank Professor Will Davies for the considerable time and effort he put into helping me with this essay. I would also like to thank all of my family and friends who have supported me greatly, proofread drafts and put up with my musings on the nature of consciousness! Balliol College, Oxford, England

References

Baars, B. J. (1988). A Cognitive Theory of Consciousness. Cambridge. Block, N. (1995). On a confusion about a function of consciousness. Brain and Behavioral Sciences 18(2), 227–247. Darwin, C. J., M. T. Turvey, and R. G. Crowder (1972). An auditory analogue of the sperling partial report procedure: Evidence for brief auditory storage. Cognitive Psychology 3(2), 255–267. Goldman, A. I. (1993). Consciousness, folk psychology, and cognitive science. Consciousness and Cognition 2(4), 364–382. Nagel, T. (1974). What is it like to be a bat? Philosophical Review 83(October), 435–50. Neisser, U. (1967). Cognitive Psychology. Appleton-century-crofts. Phillips, I. B. (2011). Perception and iconic memory: What sperling doesn’t show. Mind and Language 26(4), 381–411. Sekuler, R., A. B. Sekuler, and R. Lau (1997). Sound alters visual motion perception. Nature 385(6614), 308. Sergent, C., V. Wyart, M. Babo-Rebelo, L. Cohen, L. Naccache, and C. Tallon-Baudry (2013). Cueing attention after the stimulus is gone can retrospectively trigger conscious perception. Current biology 23(2), 150–155. Shimojo, S. (2014). Postdiction: Its implications on visual awareness, hindsight, and sense of agency. Frontiers in psychology 5, 196.

Alexander Gruen

Sperling, G. (1960). The information available in brief visual presentations. Psychological Monographs: General and Applied 74(11). Voss, J. L., C. L. Baym, and K. A. Paller (2008). Accurate forced-choice recognition without awareness of memory retrieval. Learning & Memory 15(6), 454â&#x20AC;&#x201C;459.

‘Ought Implies Can’ and the Possibility of Group Obligations isaac hadfield

T is unclear how to assign moral obligations in situations in which group cooperation is a necessary condition for successfully accomplishing some morally desirable aim. Consider, for example, a case in which someone drowning in a nearby pond can only be saved through the cooperation of four bystanders (Isaacs 2011). Based on the widely held ethical requirement that moral obligation necessitates ability, ‘ought implies can’, for a given individual bystander we cannot say that they have an obligation to perform the rescue. The ability of any individual bystander to perform the rescue is contingent on the cooperation of the others, and therefore an obligation to perform the rescue is ruled out by this ability requirement. One answer to this puzzle is that relevant groups themselves have collective obligations, and that we as individuals derive certain obligations and responsibilities from these. So, in the example at hand we might say that the bystanders collectively have an obligation to perform the rescue, and this grounds their individual responsibilities to participate. Such posited group obligations are especially important in grounding obligations in large scale collective action problems, for instance the obligations that groups have to address climate change or global poverty, where individual actors might be unable to make any difference on their own. However, positing group level obligations has come under attack from concerns relating to agency as a necessary requirement for obligation. Roughly stated, the worry is that since only agents can have moral obligations, and groups are not agents, groups cannot have moral obligations. The intuition behind this constraint is itself based on the ability requirement of ‘ought implies can’: in order for a group to have an obligation it must have the ability to perform an action, but only agents can have abilities. One prominent strategy to mitigate the force of this argument is to claim that some groups are in fact agents. This is a move particularly spearheaded by List and Pet-

British Undergraduate Philosophy Review 1 (2020)

Isaac Hadﬁeld

tit (2011), amongst others, who argue that sufficiently organized groups, for instance some corporations, are agents in their own rights and so can bear moral obligations. This paper makes a different move, and suggests that prior literature on collective obligations has ignored or too quickly dismissed work in shared agency in the philosophy of action. From accounts of shared agency we can develop a concept of joint ability, undermining the problem of agency for group obligations.

1 The agency argument According to the agency principle, only agents can have moral obligations. In cases in which we think it intuitive for groups to have moral obligations, it is not clear who exactly the agent is. Does it really make sense to say that a group can be a moral agent? If not, it does not seem like groups can have obligations. We can formulate the agency argument against groups in its simplest form as follows: Agency Argument 1 P1. Groups are not agents. P2. If groups are not agents, they cannot have obligations. (Agency Principle) Therefore, C. Groups cannot have obligations The first premise, that groups are not agents, can be subject to doubt in some cases. For List and Pettit, many sufficiently structured groups meet their preferred conditions for agency, and so can be held morally responsible. Alternatively, on Peter French’s (1984; 1995) account, groups become ‘collective agents’ when they possess a collective decision-making structure. However, even on such expansive accounts of agency like these, most groups will be excluded. In the case of the pond rescue for instance, the four previously unconnected individuals who might happen to be walking past the pond and constitute the group of bystanders do not appear to constitute a moral agent in their own right. The strategy considered in this paper aims to encompass groups like these, which we do not typically think of as moral agents. Turning now to the second premise of the agency argument, the agency principle states that only agents can have obligations. As groups are not agents, they cannot 1 Versions of this argument in a similar form can be found in Wringe (2014b:

60).

175) and (Collins 2019:

Group Obligations

have obligations. This principle is intuitive enough: to have an obligation you must be able to fulfil that obligation, and the only kind of thing that can fulfil an obligation is an agent. Since ‘ought implies can’, then as non-agential groups do not have abilities they cannot bear obligations. From this, we can construct an argument for the agency principle for groups, the ability argument: Ability Argument 2 P1. If an entity has an obligation, then that entity has an ability to fulfil the obligation (Ought implies can) P2. Groups that are not agents cannot have abilities Therefore, C. If groups are not agents, they cannot have obligations. (The Agency Principle) One challenge to the first premise comes from Wringe (2010, 2014a), who suggests that it is possible to be responsible for the obligations of other entities. This is a strange idea at first, but Wringe motivates it with reference to the everyday example of responsibility for children. We might think, for instance, that children in some situations have an obligation to behave themselves, but that it is the parents who have the responsibility to make sure the children fulfil this obligation. Similarly, a collective can have an obligation to Φ but not be able to Φ; rather, individual members are responsible for discharging Φ on the collective’s behalf. However, as Schwenkenbecher (2013) points out, this account conflates two different obligations. The first is the obligation of the children to keep quiet, or in the collective case, to address the problematic situation. The second is the obligation of the parents to keep the children quiet, or in the collective case, for members to do their part in making sure the collective addresses the problematic situation. Parents are not responsible for discharging their children’s obligations; rather, they discharge their own separate obligations towards their children. It is still the case that having an obligation implies having a corresponding ability to fulfil that obligation. If a child genuinely does not have an ability to be quiet, for instance, then we would not say they have an obligation to do so. The ability argument relies on the second premise: groups that are not agents cannot bear abilities. Yet sometimes it seems that groups do have the ability to do things. 2 This is a slightly modified version of Collins’s presentation of the argument (Collins 2019: 61).

Isaac Hadﬁeld

Lawford-Smith (2012: 463–4) poses a consideration of whether the German military had the ability to overthrow Hitler in a military coup. The German military was physically close to Hitler and had sufficient weapons and training; as such, if any group had the ability to overthrow Hitler, the German military did. For (Lawford-Smith 2012), group ability is determined with reference to group members’ abilities; individual members of the group have certain abilities, and these determine the collective ability in aggregate. This method of determination becomes clearer when we consider what it would mean for a group not to be able to do something. We might think that collective action is ruled out if enough members cannot perform their contributory action such that the minimum subset of individuals doing their part needed for the action is not met. For any individual soldier, had they attempted to do their part in overthrowing Hitler, perhaps by planning a coup, they would have been met with swift execution. As such it does not seem that the individual members of the German military did have the ability to do their part in overthrowing Hitler, and it therefore seems wrong to say that the German military as a group had the ability to overthrow Hitler. Such a formulation perhaps leaves open that groups might have abilities and thereby still have obligations in more permissive cases in which members can perform their contributory actions. Later, however, Lawford-Smith (2012) strengthens her account. What the ability argument is getting at here is just the principle that ‘ought implies can’. Furthermore, when we invoke the ought-implies-can constraint to reply to supposed obligations, by ‘can’ we do not just mean nomological or metaphysical possibility. For instance, if it would be morally desirable for me to attend a protest in London but my train from Oxford breaks down midway and I am unable to make it in time, it would be inappropriate to say that I was failing to act on my obligation just because my attending the protest would not violate any laws of nature. What ‘can’ means here concerns agency specifically, so we should ask, ‘Would the group be likely to succeed in acting if it tried?’ (Lawford-Smith 2012: 234–235). Asking this question of groups does not make sense, however; groups do not even have the functional equivalence of the relevant intentional states involved in intentional action. As groups that are not agents cannot have abilities, they therefore cannot have obligations.

2 Shared agency and joint intentionality The mistake Lawford-Smith makes here is assuming that abilities need to be held by a single agent. Many abilities are not held by any individual agent, as is obvious for

Group Obligations

actions exercized under shared agency. Consider Gilbert’s (1990) famous example of the act of two people walking together. We might go for a walk together, but neither of us individually has the ability to go for a walk together. I cannot make it the case that we go for a walk together, and similarly, you cannot make it the case that we go for a walk together; only we can make it the case that we go for a walk together. The examination of cases of shared agency has generated three broad strands of analysis. The first, championed by Gilbert (1990, 2009), is that the subject of any of these joint actions is a kind of collective agent. However, this strand would invoke a wildly permissive sense of agency, and as such Gilbert’s view tends to be rejected in the contemporary literature. The other two strands analyse cases of shared agency in terms of joint intentionality. The second accounts for joint intentionality in terms of individual intentions. This account is primarily the project of Bratman (1999, 2009, 2014), who analyses joint intention as the interlocking of individual intentions, where this cooperation is common knowledge and individual sub-plans mesh with regards to the joint activity. For example, if we were to plan a trip to New York, my sub-plan might be to book the flights, while yours might be to book the hotel. As such, Bratman’s account reads as follows: We intend to do J if and only if: 1. (a) I intend that we J and (b) you intend that we J 2. (b) I intend that we J in accordance with and because of 1a, 1b, and meshing sub-plans of 1a and 1b; you intend that we J in accordance with and because of 1a, 1b, and meshing sub-plans of 1a and 1b. 3. 1 and 2 are common knowledge between us (Bratman 1999: 121). The third strand of analysis is to posit a distinction between intending in the Imode and intending in the we-mode. What this means differs slightly between proponents. For Searle (1990, 1995), we-mode intending is a kind of psychological state, such that even a brain in a vat could intend that, for instance, we go for a walk. For Tuomela (2003, 2005), we-mode intending is a matter of intending to act as a member of a group due to having group-level reasons. A group then has a joint intention to X when (a) each agent has a we-intention to perform X and (b) there is a commonly held belief that (a) (Tuomela 2005). Rather than coming down on one side or the other, my claim here is that however we analyse joint intentions, we can extend such an analysis to give an account of joint intentional action. To return, then, to Lawford-Smith’s insistence that the problem is

Isaac Hadﬁeld

that it does not make sense to ask whether the group is likely to succeed if it tried, after consideration of these cases it now seems that it really does make sense that groups can try. The orchestra can try to play the symphony, the football team can try to win the match, and we can try and go to New York together. Accounts of shared agency like Bratman’s or Tuomela’s can be naturally extended to specify what we would mean. For instance, we might say, ‘We try to Φ just in case we act on our intention to Φ’.

3 Towards an account of joint ability Can we say something more concrete of joint abilities? One account has been provided by Pinkert (2014). The intuition predicating this account comes from consideration of cases in which there is a lack of clarity about how to coordinate. In many cases of group action, it is unclear what members are expected to do. For example, consider Pinkert’s example of the concert audience: A concert audience is awaiting a performance. Just before the performance starts, the conductor tells the surprized audience that the performance is interactive, which requires that in each row of seats, precisely half of the audience simultaneously stands up at the opening chord. (Pinkert 2014: 190) There is no obvious way for half of the individuals to jointly stand without some deliberation. It seems clear in this case that what the individuals in the audience should do is deliberate with each other to generate some procedure concerning how to organize such that half stand up. They might, for example, agree that starting from a particular end of each row, every other member of the row should stand. Sometimes a group may be immediately able to perform an action. For example, in the pond rescue case, it might be obvious what the group needs to do: each bystander needs to jump in and help save the person drowning. However, cases like the concert audience are more complicated: there is more than one salient way for the group to act, so the situation requires deliberation. Given this distinction, Pinkert devises the following conditions for an ‘immediate’ and ‘mediate’ joint ability for a group of agents ‘aa’: Immediate Joint Ability: Agents aa are immediately jointly able to Φ if and only if there is exactly one salient possible collective pattern of actions of the relevant aa-s that constitutes aa Φ-ing, and which is such that every relevant agent believes of the action which is her part in that

Group Obligations pattern that she needs to perform this action if aa is to Φ. (Pinkert 2014: 194) Mediated Joint Ability: Agents aa are at t1 mediatedly jointly able to Φn at tn if and only if 1. They are at tn immediately jointly able to perform action Φ1 at t1 , where 2. performing Φ1 at t1 makes it the case that aa are at t2 > t1 immediately or mediatedly jointly able to Φn at tn , and 3. the relevant members of aa believe at t1 that aa performing Φ1 at t1 has this effect, and that if aa is to perform Φn at tn , they will perform Φ1 at t1 (Pinkert 2014: 200).3

Pinkert’s account of joint ability is compelling, but I think that Pinkert’s approach to developing an account of joint ability around the intuitive immediate/mediate distinction is mistaken. Rather, we should extend a joint intention account of shared agency such as Bratman’s or Tuomela’s into an account of joint ability in just the same way as we would in developing an account of ability for an individual agent with individual intentions. For instance, on the conditional analysis of ability, S has the ability to Φ if and only if S would Φ if S tried to Φ (Maier 2018). Extending this analysis to a group of agents, and using Pinkert’s ‘aa’ notation, yields the following: aa have the joint ability to Φ if and only if aa would Φ if aa tried to Φ. Further, utilising the account of joint trying that I have outlined, we can therefore give a corresponding account of joint ability, the joint intention account: Joint Ability (Joint Intention Account): A group of agents aa are jointly able to Φ if and only if, given that each agent in aa acted on their intension that aa Φ, then aa is likely to succeed in Φ-ing. This formulation seems to me to be preferable to Pinkert’s, since it does not require joint ability to be anything fundamentally different from individual abilities. Furthermore, the intuitions that Pinkert aims to capture with his immediate/mediate ability distinction are also captured by my approach. Consider an example of immediate and mediate abilities for an individual. For instance, we might say that I have a 3 Pinkert combines the two into a single larger formulation but keeping them separate here helps elucidate Pinkert’s idea.

Isaac Hadﬁeld

mediate ability to make a pizza for dinner if I first need to travel to the shop to buy one. The fact that it is clear that I need to buy a pizza from the shop first means that this intermediary action is part of my trying to make a pizza for dinner. We do not need to invoke the immediate/mediate distinction when theorising about individual ability. Similarly, the joint intention account means that this distinction is unnecessary; the intuition of a group often needing to organize first to perform an action, which Pinkert is trying to capture, is implicitly allowed. If it is clear that the group needs to organize first before Φ-ing, then the group’s organising is part of the group’s acting on the intention to Φ. Furthermore, it seems that there are cases of joint ability where nobody is coordinated. For instance, Wringe (2019) gives the example of a sinking ship carrying a group of individuals. A subset of those people can swim, and by swimming to shore they can stop the boat from sinking. It may be obvious that this is what needs to be done, and so the swimmers do this without coordinating or communicating in any way. It seems clear here that the swimmers have a joint ability in just the same way as we have the ability to take a walk together. The joint intention account captures cases like this nicely, as by jumping off the ship, the swimmers are indeed acting on the intention that the group jumps off the ship such that they will prevent the ship from sinking. The simple conditional analysis of ability that I have used to develop my account of joint ability has well-known objections.4 However, the underlying point here is that whatever modifications are necessary for an account of individual ability will apply in the same way to joint ability. The main reason there has been reluctance to attribute abilities to groups is that groups are not agents and do not have the capacity for intentional action, which is intuitively necessary for abilities. Utilising accounts of shared agency and joint intentionality allows us to see that groups of agents can jointly possess abilities and that these abilities are grounded in individual accounts of ability.

4 Conclusion Group obligations are often disregarded on the basis that groups do not constitute moral agents. This paper has argued that accepting that groups of individuals can act jointly—and, as such, can have joint abilities—defeats objections to collective obligations based on the rule that ‘ought implies can’. Further, an original account of 4 One of the most famous objections to the simple conditional analysis of ability is Lehrer’s (1968) counterexample of being offered a bowl of red sweets but, having an extreme fear of red sweets, declining to take one. It might be true that if he had tried to take a sweet, he would have done. Yet this surely does not mean that he is able to take one of the sweets.

Group Obligations

joint abilities, the ‘joint intention account’ has been sketched. The utility of this manoeuvre is clear; if it is successful, then many commonplace attributions obligations become intelligible. ‘Humanity has an obligation to prevent the worst effects of climate change’ for instance could come out as true, since humans may jointly be able to accomplish this aim, rather than false from the outset since ‘Humanity’ is not a moral agent. Groups that are not agents can have abilities because they are joint abilities, shared by a group of agents. We should reject the second premise of the ability argument and as such reject the second premise of the agency argument. Groups ‘can’, and as such, groups ‘ought’. Trinity College, Oxford, England

References Bratman, M. E. (1999). Faces of Intention: Selected Essays on Intention and Agency. Cambridge, UK: Cambridge. Bratman, M. E. (2009). Shared agency. In C. Mantzavinos (Ed.), Philosophy of the Social Sciences. Cambridge, UK: Cambridge. Bratman, M. E. (2014). Shared Agency: A Planning Theory of Acting Together. Oxford: Oxford. Collins, S. (2019). Group Duties: Their Existence and Their Implications for Individuals. Oxford: Oxford. French, P. A. (1984). Collective and Corporate Responsibility. New York: Columbia. French, P. A. (1995). Corporate Ethics. Fort Worth, Tex: Harcourt Brace College Publishers. Gilbert, M. (1990). Walking together: A paradigmatic social phenomenon. Midwest Studies In Philosophy 15(1), 1–14. Gilbert, M. (2009). Shared intention and personal intentions. Philosophical Studies 144(1), 167–187. Isaacs, T. (2011). Moral Responsibility in Collective Contexts. Oxford: Oxford.

Isaac Hadﬁeld

Lawford-Smith, H. (2012). The feasibility of collectives’ actions. Australasian Journal of Philosophy 90(3), 453–467. Lehrer, K. (1968). Cans without ifs. Analysis 29(1), 29–32. List, C. and P. Pettit (2011). Group Agency: The Possibility, Design, and Status of Corporate Agents. Oxford: Oxford. Maier, J. (2018). Abilities. In E. N. Zalta (Ed.), The Stanford Encyclopedia of Philosophy (Spring 2018 ed.). Metaphysics Research Lab, Stanford University. Pinkert, F. (2014). What we together can (be required to) do. Midwest Studies In Philosophy 38(1), 187–202". Schwenkenbecher, A. (2013). Joint duties and global moral obligations. Ratio 26(3), 310–328. Searle, J. R. (1990). Collective intentions and actions. In P. R. Cohen, J. L. Morgan, M. E. Pollack, and M. Jerry (Eds.), Intentions in Communication, pp. 401–415. Cambridge, MA: MIT Press. Searle, J. R. (1995). The Construction of Social Reality. Penguin. Tuomela, R. (2003). Collective acceptance, social institutions, and social reality. American Journal of Economics and Sociology 62(1), 123–165. Tuomela, R. (2005). We-intentions revisited. Philosophical Studies 125(3), 327–369. Wringe, B. (2010). Global obligations and the agency objection. Ratio 23(2), 217–231. Wringe, B. (2014a). Collective obligations: Their existence, their explanatory power, and their supervenience on the obligations of individuals. European Journal of Philosophy 24(2), 472–497. Wringe, B. (2014b). From global collective obligations to institutional obligations. Midwest Studies In Philosophy 38(1), 171–186. Wringe, B. (2019). Global obligations, collective capacities, and ‘ought implies can’. Philosophical Studies, 1–16.

A Defence of Hilary Putnam’s Theory of Reference via Temporal Externalism alexander houghton

will demonstrate that Hilary Putnam’s theory of reference regarding ordinary natural kind terms can be preserved by a natural inclusion of temporal externalism due to its focus on the division of linguistic labour and deference to experts. Although explored in greater detail later, ‘temporal externalism’, in this instance, is essentially the claim that the meaning of certain terms at the time of their use can be determined after that time, i.e. in the future. First, I will lay out Putnam’s theory from the ground up and define terms so far mentioned. Second, I will consider Jessica Brown’s higher-level natural-kinds objection to Putnam’s theory, and articulate her own theory. Third, I will show why Brown’s account is vulnerable to a counter-example originally provided by Gary Ebbs. Fourth and finally, I will explain temporal externalism in more detail and demonstrate why it can circumvent both the obstacles presented by Brown and Ebbs, and is in fact a natural extension of Putnam’s own theory, thus preserving it against the above objections.

1 Putnam’s theory of reference 1.1 What is an ordinary natural kind term? In order to understand Putnam’s (1975) account of what determines the reference1 of ordinary natural kind terms (henceforth, ONKT), we must first examine exactly what an ONKT is. ‘Ordinary’ and ‘natural’ simply mean that what the ONKT is referring to is something naturally occurring, as opposed to some human-made artefact. A ‘term’ for our purposes simply means a 1 Also sometimes called the ‘extension’ in the literature

British Undergraduate Philosophy Review 1 (2020)

Alexander Houghton

word that refers to some object, a naturally occurring object in this case. The philosophical usage of ‘kind’ corresponds to a category/class which can be individuated from other categories/classes by certain properties which are necessary and sufficient for membership of that specific category/class. So in summary, an ONKT is a word (term) that refers to some category (kind) naturally occurring (ordinary and natural) and that can be individuated from other categories (kinds) by the necessary and sufficient properties which make it a member of that category (kind), as above. For example, ‘gold’ is a word which refers to some kind of natural objects which can be individuated from the kind of natural objects ‘silver’ refers to via their differing necessary and sufficient properties, and thus is an ONKT. The next question, then, is what sort of necessary and sufficient properties should we use to individuate objects from each other? i.e. what do we use to say gold and silver are different? In other words, what kind of kinds should be used. Putnam thinks the right kind of kind for this purpose are scientific kinds2 , as opposed to non-scientific kinds. Scientific kinds are those which are individuated by only scientific/fundamental properties (e.g. chemical structure). Non-scientific kinds are those which are individuated by non-scientific/superficial properties (e.g. physical appearance).3 Thus, according to Putnam, in trying to differentiate between gold and silver we should focus on things like their differing fundamental properties, rather than things like their colour. This has appeal when one considers examples like gold and pyrite (commonly called ‘fool’s gold’) which have similar physical appearances but differing chemical structures. This is the significance of Putnam’s theory, in tying terms to scientific kinds he is able to provide a single unambiguous and thus determinate reference, i.e. the term can only refer to one thing due to the attention to discrete scientific elements. As we shall see in later sections, deviation from the scientific method is disastrous for such a theory of reference. 1.2 How do we fix the reference of terms to scientific kinds? So far, we know what an ONKT is and that the kinds we should use are scientific kinds. The question then becomes how do particular ONKT come to refer to particular scientific kinds, i.e. how does the term ‘gold’ come to refer to the scientific kind that is element 79 Au. In other words, how is the reference fixed for an ONKT? Putnam’s answer to this question, 2 Referred to as ‘natural kinds’ or ‘substance-words’ elsewhere in the literature 3 Such as a Lockean descriptivist might wish to rely upon.

Putnam on Reference

and the crux of his whole theory, depends upon two important and linked concepts: the division of linguistic labour and a deference to experts.4 The division of linguistic labour means that not everyone in the linguistic community (i.e. language/dialect/etc.) needs to know what particular scientific kind an ONKT references to still use it correctly, and importantly for it to still refer to that scientific kind. For example, ‘gold’ will refer to the scientific kind of element 79 Au and one can still use ‘gold’ without knowing this. This is due to a deference to experts, which means that the scientific kind a particular ONKT refers to is the particular kind which experts have so far applied it to. In this way, the experts in the linguistic community bear the main burden of the division of linguistic labour. In summary, the scientific kinds ONKT refer to are determined by experts in the linguistic community, and as long as one is willing to defer to the experts one need not know what particular scientific kind a particular ONKT references to still understand that term and use it correctly within the linguistic community. 1.3 Who are the experts? The final missing piece of Putnam’s theory is determining who these experts are to whom we defer to in matters of fixing the reference of an ONKT to a particular scientific kind. Unfortunately, Putnam is rather ambiguous when it comes to identifying these experts and numerous interpretations are available. The most obvious and initially coherent interpretation is that the experts in our linguistic community we should defer to are scientific experts i.e. those with knowledge of scientific kinds (e.g. chemists, physicists, etc.). For example, ‘gold’ refers to element 79 Au because we defer to chemists, as scientific experts, because they have knowledge of scientific kinds. However, the experts-as-scientific-experts interpretation is weakened when we consider thus: we often take other’s usage of ONKT to have the same meaning and reference as our own, i.e. we take their use of words within a shared linguistic community at face value. For example, if one were holding both a bar of gold and silver and someone said ‘pass me the gold’, one would not be under any uncertainty which bar they wanted due to the presumption they are using the term ‘gold’ correctly. Importantly, this desired commonality of reference applies across time (diachronically) as well as in the present (synchronically), as we often take past usage of ONKTs within a linguistic community to be the same as our current usage. For example, we know what John Locke speaks of when he writes ‘gold’ in the seventeenth century. We 4 There is a third concept of Putnam’s theory, ‘stereotypes’.

foregoing discussion and so is left out for simplicity’s sake.

However, this is largely irrelevant to the

Alexander Houghton

take his usage of the ONKT at face value because he is part of our linguistic community. We know that by ‘gold’ he really means element 79 Au, as discussed above. Things become particularly troubling, however, when we consider that not only does Locke have no way of knowing that ‘gold’ refers to element 79 Au, but also that there are no experts to defer to in this matter since his usage of gold predates the use of differentiating elements from each other by their atomic masses (McCulloch 1995: 178). The problematic question, which we shall focus on from here on out, then becomes how can a scientifically-illiterate community (those without knowledge of scientific kinds) refer to the same scientific kind with an ONKT that later scientificallyliterate communities (those with knowledge of scientific kinds) refer to with that ONKT? How can we know that Locke really meant element 79 Au when he used the term ‘gold’? In brief, pre-scientific communities will require a different kind of expert than scientific experts, due to the latter’s unavailability. So, what sort of methods are available to a pre-scientific community when it comes to fixing the reference of an ONKT to a scientific kind? Jessica Brown’s interpretation of Putnam plausibly points to his usage of the ostensive method in pre-scientific communities, outlined below (Brown 1998: 278). Also of interest is the descriptive method found in the work of Saul Kripke in his own theory of reference (Kripke 1980). The ostensive method involves simply pointing at some objection and saying something along the lines of ‘that is called xyz’. This would thus fix the reference of ‘xyz’ to whatever scientific kind is instantiated in that object.5 For example, pointing at a lump of gold and calling it ‘gold’ would fix the reference of that term to element 79 Au. The descriptive method involves calling a number of objects by some term ‘xyz’. This would fix the reference of ‘xyz’ to whatever scientific kind those things have in common. For example, calling both fresh water and salt water ‘water’ would mean that term refers to the scientific kind they share, namely h20. An initial appeal of both of these approaches is that they would be able to preserve commonality of reference into a more scientifically-literate community. However, as we shall shortly see, there is a grave problem with both the ostensive and descriptive methods of reference-fixing available to a pre-scientific Putnamian community. 5 Kripke (1980: 135) refers to this as the object’s ‘baptism’.

Putnam on Reference

2 The higher-level objection 2.1 The problem with the ostensive and descriptive methods This problem for the ostensive and descriptive methods is Jessica Brown’s higher-level natural-kinds objection (Brown 1998: 279). The problem, Brown notes, is that ‘typically, an item that instantiates one natural [scientific] kind instantiates several’ (Brown 1998: 275), and neither the ostensive nor descriptive method of reference-fixing can distinguish between these multiple scientific kinds. To understand why something which instantiated one scientific kind will instantiate several scientific kinds, consider the case of gold once more. Numerous kinds are instantiated in any naturally occurring gold: element 79 Au as mentioned above, the more general kind that is metal, and a variety of isotopes of element 79 Au depending on the particular sample(s) of gold present. Which one of these particular scientific kinds does the term ‘gold’ refer to? If one is unable to distinguish between these different instantiated kinds then terms like ‘gold’ might end up referring to other metals like silver due to their shared scientific kind, metal-ness. As mentioned above, unfortunately neither the ostensive nor descriptive methods of reference-fixing are able to distinguish between these different instantiated kinds. The ostensive method of pointing at a lump of gold could end up fixing the reference of ‘gold’ to simply metal or a highly specific isotope of element 79 Au which would lead to some things being called gold that clearly are not, like silver, or some things which clearly are gold failing to meet the criteria, like other isotopes of gold besides the present example. The descriptive method falls prey to the same problem; trying to fix the reference of a term via their shared scientific kinds could easily once more have silver objects being called ‘gold’ for their metal-ness and cause other plausibly gold objects from being excluded from the term if they were of a different isotope. In brief, both the ostensive and descriptive methods are too ambiguous in terms of which scientific kind they are trying to pick out and tie to a term. The upshot of this is that these methods of fixing the reference of ONKT are indeterminate and unable to tie a single scientific kind to a term Thus, they have lost the main advantage of using scientific terms in the first place— a lack of ambiguity. As such, it is unclear why we should defer to experts, at least nonscientific experts, as required in Putnam’s theory. This in turn makes commonality of reference across time far less certain and would thus prevent us taking pre-scientific communities at face-value, despite our intuitive desire to do so.

Alexander Houghton

2.2 Recognitional capacities as an alternative Brown attempts to save the notion that an ONKT’s reference can be fixed by a scientific kind pre-science via the notion of recognitional capacities. To have a recognitional capacity is to have the ability to distinguish a scientific kind from other local kinds by non-scientific features. This allows one to distinguish between the different scientific kinds instantiated in a single object or a collection of objects. For example, even though gold instantiates multiple scientific kinds, as above, one can still distinguish between gold and silver by other non-scientific features (colour etc.) even though they share the scientific kind of metal-ness. Thus, the term ‘gold’ is fixed to whatever scientific kind gold instantiates that is not instantiated in other local objects. This allows for an allegedly determinate fixing of the reference of terms to specific scientific kinds pre-science. In other words, commonality of reference will be preserved across time through pre-scientific communities to scientific communities, solving the higher-level objection levelled at Putnam’s pre-scientific ostensive and descriptive methods. So rather than appealing to either non-existent scientific experts or unsatisfactory ostensive/descriptive experts, (i.e. ordinary members of the linguistic community) one can defer to experts with recognitional capacity. Such recognitional capacity experts are those with wide knowledge of the non-scientific features of objects who can thus distinguish between local kinds. Brown’s theory also requires that one must have the metaphysical understanding that what definitively determines an object’s membership of a scientific kind is its scientific/fundamental properties, not these non-scientific features. However, this is primarily to avoid the criticisms of Lockean descriptivism and thus we need not dwell on this condition in this article. In summary, the recognitional capacities approach allows Brown to defeat the higher-level objection, as she can use non-scientific methods (i.e. those available to scientifically-illiterate communities), namely recognitional capacities, to distinguish between scientific kinds whilst avoiding descriptivism by acknowledging these prescientific references are defeasible by later scientific knowledge of something’s fundamental properties. This allows Brown to attach a determinate reference to an object that will roughly align with later scientific kinds discovered by science as she can distinguish between kinds due to their non-scientific qualities. Thus, defeating the higher-level objection.

Putnam on Reference

3 The branching twin-earth objection 3.1 The thought experiment However, it would appear Brown’s theory is subject to serious doubts as well. This is demonstrated by a thought experiment presented by Gary Ebbs, which shows that recognitional capacities cannot be what allows scientificallyilliterate communities to speak of the same kinds with ONKTs as scientifically-literate ones do (Ebbs 2000: 248). The thought experiment is as follows: there are two planets that are extremely distant from each other and have no interaction, Earth and Twin-Earth. They are identical in all aspects not distinguished here. On both Earth and Twin-Earth the term ‘gold’ is used identically up until 1650. Then on Twin-Earth, what we on regular Earth call ’platinum’ is discovered. This discovery predates any molecular chemistry, and so when non-scientific tests are performed on this Twin-Earth platinum it appears identical to gold. This leads the Twin-Earthians to extend their term ‘gold’ to include this new material, even after the invention of molecular chemistry which allows them to see gold and platinum are really two separate scientific kinds. Compare this example to the discovery that gold has two separate isotopes, but even after this discovery we still call both ‘gold’. Meanwhile, platinum is discovered post-discovery of molecular chemistry on regular Earth and thus is known to be a distinct scientific kind and accorded the name ‘platinum’ as such. In short, post-1650, ‘gold’ refers to gold and platinum on TwinEarth, whilst it refers simply to gold on regular Earth due to this temporal branching. It would appear on Twin-Earth that ‘gold’ is referring to a more general scientific kind like metal or gold-coloured metals, whilst on regular Earth it has the usual referent of element 79 Au. Thus, it would appear the same ONKT ‘gold’ refers to two different things on Earth and Twin-Earth, despite both being scientifically-literate societies. 3.2 The problem for the recognitional capacities approach This causes problems for Brown’s account as there seems little reason to suppose the Earthian and Twin-Earthian’s recognitional capacities for gold differed before the discovery of platinum. The use of the term ‘gold’ would have picked out exactly the same thing on both planets, namely element 79 Au. In terms of recognitional capacities, they would have been able to distinguish gold from other local scientific kinds (e.g. silver etc.) by non-scientific features just as well on Earth as on Twin-Earth. Yet, post-discovery of platinum, their use of ‘gold’ refers to two separate kinds, gold alone on Earth and gold-and-platinum on

Alexander Houghton

Twin-Earth. This points to a grave problem for the recognitional capacities approach: it cannot be recognitional capacities that allow scientifically-illiterate communities to refer to the same scientific kinds as scientifically-literate communities do with ONKT. This is because recognitional capacities too fail to distinguish between multiple scientific kinds instantiated in one object, as per Brownâ&#x20AC;&#x2122;s own higher-level objection. Thus, the recognitional capacities theory also does not provide a determinate reference, the same problem Brown found with the ostensive and descriptive methods of reference-fixing. Thus far we have established two major and linked problems. Firstly, that the methods employed by a pre-scientific Putnamian community on one interpretation, the ostensive and descriptive methods, are inadequate to provide a determinate reference that will allow continuity of reference into scientific communities. Secondly, this problem is not resolved via Brownâ&#x20AC;&#x2122;s recognitional capacities account as it too provides indeterminate references when subjected to the branching twin-earth objection.

4 A temporal externalism rescue 4.1 What is temporal externalism and how does it avoid the above objections? So, can we explain how scientifically-illiterate communities reference scientific kinds without falling prey to either objections? Fortunately, yes, via the introduction of temporal externalism: Temporal Externalism: a term may have a particular meaning at time t in virtue of facts of its usage after time t. (Collins 2006: 58). In essence, the reference of terms need not necessarily be fixed at the moment of those termsâ&#x20AC;&#x2122; usage, especially in pre-scientific communities as we shall see below. Firstly, this concept does not fall prey to the Ebbsian counter example as it sidesteps the requirement for a determinate reference to be temporally contemporaneous with recognitional capacity/usage/etc. The key to reconciling temporal externalism with this thought experiment is to hold that the referents of both Earths were different before the discovery of platinum, not only after. These referents are different due to the differing treatment of gold and platinum due to temporal externalism retroactively delimiting the reference of the term. This allows it to avoid the difficulty presented to Brown in explaining the difference of reference post-discovery and the identical usage pre-discovery. To the temporal externalist, there was never any difference pre-/postdiscovery. Rather, a reference may be made determinate by later events.

Putnam on Reference

Secondly, neither does temporal externalism fail due to Brown’s higher-level objection for similar reasons. Indeed, any instantiation of an object will present multiple kinds to a pre-scientific community, however, the temporal externalist rejects any requirement to give a determinate reference in that moment of scientific-illiteracy. Note this does not prevent the temporal externalist from holding the intuitive claim of commonality of reference across time, as mentioned above, i.e. that scientifically-illiterate communities refer to the same scientific kind as scientifically literate ones. 4.2 How does temporal externalism save Putnam? Most importantly temporal externalism can answer our original question by explaining how scientifically-illiterate communities speak of scientific kinds through the division of linguistic labour and deference to experts. As Collins notes, ‘scientifically-ignorant communities participate in the same temporallyextended linguistic practice’ (Collins 2006: 59) as later scientifically-literate societies. Why would a deference to scientific experts not extend temporally as well? This is what allows the temporal externalist to avoid any charges of indeterminate reference by shrugging the question off to future, more scientifically-literate communities engaged in the same linguistic practice as their pre-scientific ancestors. In short, scientificallyilliterate communities may refer to kinds with ostensive/descriptive means as long as they recognize these are defeasible by later scientific knowledge. Things begin to look strikingly Putnamian when we consider the role of division of linguistic labour and deference to experts in temporal externalism. In fact, there is already a temporal delay in ordinary division of linguistic labour and deference to experts, for surely any scientific information is not transmitted immediately to the scientifically illiterate. Thus, Putnam’s account can be saved by the natural extension of his concept of division of linguistic labour into temporal externalism via deference to future possible scientific experts. 4.3 Is temporal externalism plausible? So, is this temporal externalism plausible? One initially alarming attack is the accusation that temporal externalism implies backwards causation. Collins rejects this, demonstrating the relation temporal externalism reflects is not casual but truth-making (Collins 2006: 60). For example, take the phrase: ‘George W. Bush becomes the forty-third President of the United States’ (Collins 2006: 60) being uttered at the beginning of his presidential campaign. Such a statement plausibly did not have a truth-value at the moment of its utterance, for

Alexander Houghton

it spoke of future events. However, it can be said that the statement has been made retroactively true by how history turned out. Whilst a greater discussion of the objections to temporal externalism are in order, it is unfortunately outwith the scope and length of this article. However, it is worth noting Collins claims that: ‘most criticisms of TE [temporal externalism] are only particular instances of criticisms of content externalism in general’ (Collins 2006: 60). Thankfully, others have already given temporal externalism an excellent defence in our stead.6

5 Conclusion In conclusion, initially it appears Putnam’s theory of reference is undermined by Brown’s higher-level objection. However, Brown’s own account is undermined by Ebbs’ counter example. The only way to save any of the above theories of reference is to accept that division of labour and deference to experts is temporally extended. For reasons shown above, this is not an implausible step and is a natural progression of Putnam’s own community focused theory. Unfortunately, Brown’s theory fails to survive this process due to a lack of focus on the division of linguistic labour, although Brown does come close due to her acknowledgment of determinate reference coming finally through science. Acknowledgements The author would like to thank Erika Foster and Lesley Affrossman for their support throughout. The University, Glasgow, Scotland

References Brown, J. (1998). Natural kind terms and recognitional capacities. Mind 107(426), 275–303. Collins, J. M. (2006). Temporal externalism, natural kind terms, and scientifically ignorant communities. Philosophical Papers 35(1), 55–68. Ebbs, G. (2000). The very idea of sameness of extension across time. American Philosophical Quarterly 37(3), 245–268. 6 For example, (Jackman 1999; Tanesini 2014).

Putnam on Reference

Jackman, H. (1999). We live forwards but understand backwards: Linguistic practices and future behavior. Paciﬁc Philosophical Quarterly 80(2), 157–177. Kripke, S. (1980). Naming and Necessity. Harvard. McCulloch, G. (1995). The Mind and its World. Routledge. Putnam, H. (1975). The meaning of ’meaning’. Minnesota Studies in the Philosophy of Science 7, 131–193. Tanesini, A. (2014). Temporal externalism: A taxonomy, an articulation, and a defence. Journal of the Philosophy of History 8(1), 1–19.

A Brief Conceptual Analysis of what ‘Implicit Attitude’ Means milena koot

1 Introduction REENDWALD & Banaji, inventors of the Implicit Associations Test (IAT), define implicit attitudes as ‘introspectively unidentified (or inaccurately identified) traces of past experience that mediate favourable or unfavourable feeling, thought, or action towards social objects.’ (1995, p. 8). I want to dig deeper into what this means and object to what often amounts to a dual-system account of implicit attitudes. Section 1 of this paper shows how philosophers and psychologists often employ dual-system thinking (Tversky & Kahneman, 1974), such that implicitness is equated with System 1 characterised by a bundle of features such as unintentional, efficient and automatic. Section 2 criticises uses of dual-system theory and concludes that this framework inadequately explains implicitness. In response to this, section 3 discusses the decompositional view of automaticity applied to implicit bias. It shows that implicit biases respond to situational constraints and lack a consistent bundle of features. This raises the possibility that implicit bias is not a natural kind. Rather, implicit bias measures capture reactions to certain situational constraints.

2 Implicit bias and dual-system theory Conceptions of what the term ‘implicit’ means are influenced by dual-system theorising: Kahneman (2011) and Evans (2008) posit two distinct processing systems. System 1 processing is unconscious, automatic, fast and associative, while System 2 processing is conscious, deliberate, slow, and propositional. These features are similar to those defining implicit bias measures like the IAT (for IAT procedure details see section 3). Here implicit bias is conceptualised as characteristically unintentional, efficient, automatic, unconscious, and associative. Therefore, if we are asking what ‘implicit’ means, British Undergraduate Philosophy Review 1 (2020)

Implicit Attitude

it is often suggested that a process is implicit if it has System 1 features. For example, both the ‘Associative-Propositional Evaluation’ (Gawronski & Bodenhausen 2011) and ‘Reflective-Impulsive Model’ (Strack & Deutsche 2004) conceptualise implicit bias as part of an impulsive system based on contiguous and contingent associative learning. On this theory, associative networks generate spontaneous valence responses to the world. These two models imply an ‘alignment view’ of implicitness—defining ‘implicitness’ via a bundle of features: associative, impulsive, spontaneous, unconscious, and non-deliberate. Psychology is famous for dual-processing approaches, but philosophers frequently use them too. Corneille & Stahl (2019) point out that dual-system models are ‘consistent with a long attitude research tradition tracing back to Plato, (which) opposes affect-based to cognition-based learning (e.g. Edwards 1990).’ (p.178). Fridland (2017a) also illustrates philosophers using this dichotomy: Stanley (2011) in Know How, Di Nucci (2013) in his book Mindlessness, and Dreyfus in the paper Intelligence without representation (2002), all posit dichotomies of ‘intelligent, rational, conscious, controlled processes on one hand and unintelligent, automatic, passive, unconscious routines on the other.’ (p.4341). This theorising extends to philosophical models of implicit bias. For example, Tamar Gendler posits a sui generis mental state called ‘Alief’ to characterise implicit bias, to distinguish it from belief. Aliefs are characterised as ‘associative, automatic and arational. They are also affect laden and action generating’ (2008b, p. 557 original emphasis). Aliefs are acquired through habit or exist innately. Gendler’s view is action-oriented, as Aliefs are iterated by responding to a stimulus in a particular way (automatically and arationally). Aliefs capture a number of contradictory belief occurrences including phobias, fictional emotion, and bad habits (Gendler, 2008b, p. 554). This model implies an ‘alignment view’ of implicitness, defining ‘implicitness’ via a bundle of features. The features of associative, irrational, and automatic come in a bundle and align in the class of Aliefs and are distinct from the features aligned for the class of beliefs. The Alief theory, Associative-Propositional Evaluation and ReflectiveImpulsive Model, are also popular, exemplifying the reach of the alignment view. Theoretical approaches involving alignment often infer the presence of one feature from the presence of others (Melnikoff & Bargh, 2018). For example, believing that certain features align into a natural kind, researchers infer the presence of ‘unconscious’ from the presence of a feature such as ‘unintentional’. Philosophers and psychologists alike use dual-system approaches, to equate implicit processing with a bundle of System 1 features. The next section critically exam-

Milena Koot

ines whether this approach can provide a satisfying conceptual analysis of ‘implicit’.

3 Against a dual-system approach to implicit bias This section provides four objections to viewing the term ‘implicit’ via a dual-system approach. Firstly, implicit bias can be introspectively accessed: Hahn and colleagues (2014) showed that participants who never encountered implicit bias testing, could, upon reflection, give an accurate prediction of their IAT score. This prediction was made independent of explicit biases, which diverged from implicit scores. This suggests implicit bias can be available to introspection. This contradicts the dual-system view of implicitness, because System 1 processing is characteristically unconscious. Secondly, implicit bias is influenced by explicit processing. A second strain of research suggests that implicit bias is sensitive to conscious thought. Namely, debiasing evidence suggests implicit bias is affected by information stored in propositional format. For example, participants were presented with a negative but true story from Mahatma Gandhi’s life, and this propositional learning changed participants’ implicit bias towards Gandhi (Van Dessel, 2017). Byrd (2019) criticises how propositional debiasing is often taken as evidence that implicit bias is propositional in nature. While a mental process is either associative or propositional, a behaviour or attitude can be based on both associative and propositional processes. For example, part of what makes most people pick their toothbrush up in the morning may be an association between morning and toothbrush. But similarly, a propositional intention such as ‘In the morning, I brush my teeth’ can simultaneously contribute to the behaviour. Therefore, a behaviour can be grounded in both associative and propositional processes. This is supported by Devine et al. (2012) who find that reflective counterconditioning can change implicitly biased behaviour. This suggests implicit bias likely has explicit as well as implicit mechanism underlying it, because explicit instruction can affect implicit bias scores (Byrd, 2019).1 Therefore, debiasing exceeds the features characteristic of System 1. This shows that implicit processing does not exist in an entirely separate realm from explicit processing. Thirdly, the implicit bias feature ‘fast’ does not consistently align with features 1 An associativist may argue this evidence is compatible with the view that propositional knowledge can actually be reduced to associative knowledge. But the debate about whether the mental states underlying implicit bias are associative or propositional is tangential to this paragraph’s point. The information being processed explicitly in the debiasing experiment is reflective and therefore shows features of System 2. Therefore, implicit bias not only operates in System 1, but is affected by System 2, regardless of whether the mental state underlying implicit bias can be reduced to associative or propositional knowledge.

Implicit Attitude

‘uncontrolled’ or ‘unintelligent’. Fast and automatic processing can be controlled and logical. Bago and De Neys (2016) test the time course assumption of dual-system theories, namely that System 1 is fast while System 2 is slow. They show that quick, fast, automatic responses were more logical than expected. Speed is usually characteristic of System 1 while logic is characteristic of System 2, however, these features overlapped. This is not uncommon. An illustrative example is expert chess playing. A chess champion can make strategic moves at top speed and based on mere pattern recognition. But the moves are logical, strategic, sensitive to the situation, and based on skilful practice, while still being fast and automatic. Therefore, speed and automaticity do not necessarily entail lack of control or intelligent responding. This evidence contradicts the dual-system assumption that fast and automatic processing is independent of control or logic. Fourthly, evaluative conditioning (which very likely contributes to implicit bias) is moderated by goals. Evaluative conditioning may adjust according to goals, intention, and actions of the agent. For example, Corneille and Stahl (2019) provide a thorough review and find that evaluative conditioning effects are sensitive to processing goals, as facilitated by experimental set-up. When the experimental set-up facilitates focus on associative pairing or on valence (Gast & Rothermund, 2011b) then larger evaluative conditioning effects are found (Kattner, 2012). On the other hand, when the set-up focuses on redundant task features (such as brightness) rather than relevant dimensions of the stimuli (such as valence) a smaller effect occurs (Stahl et al., 2016). When the goal is to process stimuli separately rather than as a pair, then even smaller evaluative conditioning effects occur (Stahl et al. 2016). In conclusion, evidence does not support a System 1 view of implicit bias. This section shows that ‘implicit’ cannot be understood in terms of the bundle view, according to which features of speed, automaticity, association, and unconsciousness always go together and characterise an implicit process. However, this is still how implicit/explicit dichotomies are frequently perpetuated in introductory cognitive science lectures and textbooks. Therefore, the question arises, how should implicit bias be defined if the bundle view does not succeed? The next section discusses a decompositional approach to implicit bias. A decompositional approach has traditionally been used for conceptual analysis of the term ‘automatic’. Moors, Spruyt & De Houwer (2010) also suggest this approach for defining implicit measures, in the context of praxis. But this approach has not been philosophically applied, to my knowledge, in a direct analysis of the meaning of implicitness.

Milena Koot

4 A decompositional approach to implicit bias The decompositional approach originally comes from research into automaticity. I give a very brief overview of it. Two pieces of evidence speak in favour of a decompositional approach to automaticity: a lack of conceptual overlap between automaticity features and a lack of one-to-one relation among automaticity features (Moors & De Houwer, 2006a). Features of automaticity are: goal-relatedness, consciousness, efficiency, and fastness. These features can be separated, both conceptually and experimentally, contrary to the bundle approach discussed in the previous section. Unfortunately, there is not the scope here to justify the decompositional approach to automaticity in full but to read about how automaticity features are conceptually distinct and de facto come apart, see Moors & De Houwer, (2006a). Most importantly, the decompositional approach to automaticity implies the following improvement. When using the word automatic to describe a process, one should specify whether this means it occurred under conditions: controlled (goalrelated) and/or unconscious and/or efficient and/or fast. Classifying an automatic process in terms of the features that facilitate it will allow researchers to investigate whether non-manipulated features are necessary (Bargh, 1992). This will justify generalising from the lab to the outside world. For example: If I specify that a racial bias occurred in the lab automatically, and by automatic, I mean: ‘it was facilitated by a condition of strong time constraints’ this justifies generalization. It means one can generalise to real world scenarios with matching time constraints. Another theoretical implication of the decompositional approach is that automaticity features are external to the automatic process. They are background conditions. Nonetheless, they are what qualify the process as automatic. Next I address how automaticity features (goal-direct, conscious, efficient, fast) apply to the IAT procedure. Implicit bias responses occur due to the specific constraints of the IAT context. Implicit bias is not rigid or fully formed and then retrieved. Rather, implicit bias is a response to a specific set of situational constraints. To begin with a brief explanation of the Race IAT.2 The initial two stages of the Race IAT are typically easy. First, participants quickly categorize a list of word (like ‘lazy’ or ‘joy’) as GOOD or BAD. They indicate their responses on computer keyboards, with different keys connoting GOOD versus BAD. In the second stage, participants categorize peoples’ photographs according to race, as either WHITE or 2 Other themes tested via IAT include sexuality, disability, age, body weight (Charlesworth & Banaji,

2017), suicide (Glenn et al., 2017), alcohol consumption (Lindgren et al., 2016).

Implicit Attitude

BLACK. In the key part of the experiment, participants are instructed to quickly pair stimuli according to a particular category rule. For example, the rule may be ‘If the face is BLACK categorize it as GOOD’ and vice versa for WHITE faces. The results show that participants more quickly and more accurately categorize White faces as GOOD as opposed to Black faces. Similarly, participants more quickly categorize Black faces as BAD in comparison to White faces (Greenwald et al., 1998). Firstly, how does the feature ‘goal-directed’ apply to the IAT? There are two different kinds of goals: goals to initiate a process versus goals to alter or stop a process. Promoting goals in the context of the IAT procedure occur in two different ways. Firstly, the IAT instructions may determine a promoting goal. If a specific session involves instructions to pair BLACK and GOOD, then this is the promoting goal. However, despite desiring to pair BLACK and GOOD, on certain trials this promoting goal is not causally efficacious, and participants fail to be goal-directed. If we interpret promoting goals this way, then sometimes participants are not goal-directed in an IAT. However, we can interpret promoting goals in a different way, for example, such that the participants operate under the promoting goal to ‘evaluate stimuli’ generally. The first section of the IAT requires words like ‘lazy’ and ‘joy’ to be categorized as GOOD or BAD. Here, the general goal of evaluation is salient. Despite the next instructions to mechanically pair BLACK and GOOD without any active evaluation, participants are primed with the goal of evaluating stimuli. Even though participants are instructed to apply the instructions dogmatically and should not use evaluation, they may use evaluation nonetheless, because evaluation has been a salient goal throughout the procedure. Similarly, polarization is encouraged because stimuli in both categories—words and faces—are extreme. For example, in the Age IAT there are only very young and very old faces but no face that is middle-aged. Similarly, the words that appear are either extremely positive or extremely negative. Therefore, a promoting goal that occurs could be polarization. Without polarization being encouraged by the experimental procedure the same IAT pattern may not occur. A number of studies support this view (Spruyt et al. 2007). They show that in one implicit bias measure (an affective priming procedure) ‘affective priming effects are only found when valence is made salient by some characteristic of the procedure (target task, stimulus onset, prime task, context).’ (Moors, Spruyt & De Houwer, 2010 p.30). In other words, task procedures make valence salient and may produce a goal to see things in a valenced way. Therefore, in one sense IAT responses are not goal-related (via the instructions), but in another sense they are (via a goal to evaluate).

Milena Koot

Secondly, how do ‘counteracting goals’ operate in an IAT? Whether counteracting goals are realized largely depends on other features such as time and attention. Counteracting goals fail when response times are forced to be short, which here means less than 600 milliseconds (Degner, 2009). Moreover, participants with significantly better-than-average working memory (and therefore attentional capacities) are good at implementing counteracting goals. Participants with significantly worse-than-average working memories are worse at implementing counteracting goals (Klauer et al., 2007). This suggests attention plays a role in control and goal-relatedness. The fact that time and attention are constraints could explain the fact that implicit bias does not predict behaviour (Forscher et al., 2019). The IAT has temporal and attentional constraints in a way ecologically valid behaviour normally does not. IATs may fail to predict behaviour because in an ecologically valid context a counteracting goal stops the process. This is supported by the fact that self-harm and suicide IATs are valid at predicting suicidal behaviour (Glenn et al., 2017). While the IAT is normally bad at predicting behaviour, in some circumscribed cases such as suicide and self-harm, it consistently predicts behaviour. This might be explained as follows: an ecologically valid suicide context likely lacks substantial cognitive attention, because mental health decline correlates with attentional difficulties and limited cognitive processing. Therefore, an ecologically valid suicide context is often similar to an IAT context in that attention is not functioning at full capacity. This is to highlight that, as the decompositional approach predicts, implicit bias is very bound by constraints of specific contexts. Thirdly, how does the feature ‘conscious’ operate in an IAT? During the IAT, participants are conscious of input (for example, a face or word). However, especially in inaccurate trials, participants can lack conscious awareness of transition that leads from input to output. In accurate cases, it is likely that participants consciously apply task instructions. In such a trial, participants are likely conscious of input, output and the transition. However, because the IAT is a forced choice task, the choice may be active or conscious in a loose sense. Participants are not conscious of all qualities of their response, for example, the speed of their response. Furthermore, there is evidence that some qualities of implicit bias fulfil the criteria for consciousness, at least for introspective availability. Recall Hahn et al.’s (2014) experiments where people accurately predicted their implicit biases. This suggests that one’s own level of bias—for example, difference in speed of responding—could be introspectively accessible. In summary, the input and output of the IAT (e.g. the stimulus and the forced choice result) are con-

Implicit Attitude

scious. While not all features of the process may be accessible (such as exact speed), people can have some general introspective access to predict their own implicit bias. Fourthly, how does the feature ‘efficient’ operate in an IAT via attention? According to Shiffrin (1988), a process counts as efficient if it operates without using substantial attention. As explained before, Klauer et al. (2007) show that having significantly better-than-average and worse-than-average attentional capacities predicts counteracting goal efficacy. Therefore, attention plays a critical role in IAT performance constraints. Lastly, how does the feature ‘fast’ operate in an IAT paradigm? Firstly, fast is gradual and relative. If responses in one condition are faster than in another condition, this condition is labelled ‘fast’. Allegedly, an unexpected pairing produces slower response times, and if the unexpected pairing were attempted quickly, there would be less accuracy. This illustrates how automaticity features in an IAT paradigm are additive. An unexpected pairing can be accommodated by more time. In other words, if the situational constraint of time did not exist, then the IAT accuracy scores would differ. Throughout this section, time arises as an important situational factor that facilitates implicit bias. In summary, during an IAT participants may fail to be goal-directed with regards to task instructions, but may be goal-directed with the general goal of evaluating, which is salient in the task design. Secondly, counteracting goal success largely depends on attention and time constraints. Conscious awareness may not apply to all features of the process (such as exact speed), but there may be introspective access to some IAT qualities. Attention, which varies across individuals strongly affects other features like goal-directedness. The same goes for time. In conclusion, these considerations lead to the following decompositional approach to implicit bias: a process is implicit if it operates under suboptimal conditions. The suboptimal conditions vary by context. They can include lack of attention, and/or time, and/or conscious stimulus input, and/or goal-directedness. This suggests the possibility that ‘implicit bias’ is not a natural kind. There is no natural alignment of features to characterize implicit bias. Rather, implicit bias is strongly context reactive and condition dependent. In procedures like the IAT, we may not be studying an ontological mental state.3 3 The ethical considerations of implicit bias are outside the scope of this essay. But suffice to say that the idea of ‘suboptimal processing’ is not meant to excuse implicit bias or at all exculpate the implicitly biased agent from blame or responsibility.

Milena Koot

5 Conclusion In this paper I started by showing how philosophers and psychologist often employ dual-system thinking such that implicitness is equated with System 1 and characterized by a bundle of features such as unintentional, efficient and automatic. I criticized uses of dual-system theory and concluded implicitness is inadequately explained by this framework. In response to this, I defended a decompositional view of implicit bias as not determined by a fixed set of features but varying by context. Implicit bias is suboptimal processing as a result of time constraints, and/or lack of efficiency, and/or lack of consciousness and/or lack of control. Therefore, researchers are encouraged to specify the particular processing conditions which produced implicitness. This analysis suggests the possibility that implicit processing is not a natural kind, as there is no natural alignment of features that characterize it. Author’s note This paper is an excerpt of a longer bachelor thesis conceptually analysing implicit bias—unpublished. Dedication This paper is dedicated to Hal Milne. Harris Manchester College, Oxford, England

References* Amodio D. & Devine P., 2009, ‘On the interpersonal functions of implicit stereotyping and evaluative race bias: Insights from social neuroscience’ in Attitudes: Insights from the new wave of implicit measures, ed. R. Petty, R. Fazio, & P. Briñol, New Jersey, Erlbaum. Allport, G.W., 1954, The nature of prejudice, Cambridge, Mass., Addison-Wesley. Asch, S.E., 1951, ‘Effects of group pressure on the modification and distortion of judgments’ in Groups, leadership and men, ed. H. Guetzkow, Pittsburgh, PA, Carnegie Press. * Due to the number of works cited, it has proven impractical to ensure the reference list conforms to house style in all particulars.—Ed.

Implicit Attitude

Arcuri L., Castelli L., Galdi S., Zogmaister C. & Amadori A., 2008, ‘Predicting the Vote: Implicit Attitudes as Predictors of the Future Behavior of Decided and Undecided Voters’, Political Psychology, Vol.29, No. 3, pp. 369-387. https://doi.org/10.1111/j.14679221.2008.00635.x. Baars, B.J., 1997, In the Theater of Consciousness: The Workspace of the Mind, Oxford, Oxford University Press. Baars, B.J., 2017, ‘The Global Workspace Theory of Consciousness’ in The Blackwell Companion to Consciousness, ed. S. Schneider & M. Velmans, Chichester, John Wiley & Sons Ltd. doi:10.1002/9781119132363.ch16. Bago, B. & De Neys, W.,2017, ‘Fast logic?: Examining the time course assumption of dual process theory’, Cognition, Vol.158, No. 90. doi:10.1016/j.cognition.2016.10.014. Bargh, J.A., 1989, ‘Conditional automaticity: varieties of automatic influence in social perception and cognition’ in Unintended Thought, ed. J.S. Uleman & J.A. Bargh, New York, The Guilford Press. Bargh, J.A. ,1992, ‘The ecology of automaticity: toward establishing the conditions needed to produce automatic processing effects’ (Special Issue: Views and Varieties of Automaticity), American Journal of Psychology, Vol.105, No.2, pp. 181-199. Bargh, J.A., 1994, ‘The four horsemen of automaticity: intention, awareness, efficiency, and control as separate issues’ in Handbook of Social Cognition Vol. 1, Basic Processes, 2nd edn., ed. R.S.Wyer, & T.K. Srull, Hove, England, Psychology Press. Blair, I. V., Ma, J. E., & Lenton, A. P., 2001, ‘Imagining stereotypes away: The moderation of implicit stereotypes through mental imagery’, Journal of Personality and Social Psychology, Vol.81, No.5, pp. 828-841. Brownstein, M., 2018, The Implicit Mind: Cognitive Architecture, the Self, and Ethics, Oxford, Oxford University Press. Brownstein, M., 2019, ‘Implicit Bias’, The Stanford Encyclopedia of Philosophy, ed. E.D. Zalta, URL = <https://plato.stanford.edu/archives/fall2019/entries/implicitbias/>. Brownstein, M. & Madva, A., 2012a, ‘Ethical Automaticity’, Philosophy of the Social Sciences, Vol.42, No.1, pp. 68-98. Brownstein, M. & Madva, A., 2012b, ‘The Normativity of Automaticity’, Mind & Language, Vol.27, No.4, pp.410-434.

Milena Koot

Brownstein, M., Madva, A. & Gawronski, B.,2019, ‘What do implicit measures measure?’, WIREs Cognitive Science, Vol.10, No.5, pp. 1501. https://doi.org/10.1002/wcs.1501. Byrd, N., 2019a, ‘What we can (and can’t) infer about implicit bias from debiasing experiments’, Synthese. https://doi.org/10.1007/s11229-019-02128-6. Carruthers P.,2013, ‘On Knowing Your Own Beliefs: A Representationalist Accoun’ in New Essays on Belief, ed. N. Nottelmann, London, Palgrave Macmillan. Cervone, D., Caldwell, T. L., & Mayer, N.D., 2015, ‘Personality systems and the coherence of social behavior’ in Theory and explanation in social psychology ed. B. Gawronski & G. V. Bodenhausen, New York, The Guilford Press. Chopik, W.J. & Giasson, H.L., 2017, ‘Age Differences in Explicit and Implicit Age Attitudes Across the Life Span’, The Gerontologist, Vol. 57, pp. S169-S177. Clark, A., 1997, Being there: putting brain, body, and world together again, Cambridge, Mass, MIT Press. Csibra, G., 2008, ‘Goal attribution to inanimate agents by 6.5-month-old infants’, Cognition, Vol.107, No.2, pp.705-17. Csibra, G. & Gergely, G., 2009, ‘Natural pedagogy’, Trends in cognitive sciences, Vol. 13, No. 4, pp. 148-153. Cohn, A.M., Cameron, A.Y., Udo, T., Hagman, B.T., Mitchell, J., Bramm, S. & Ehlke, S., 2012, ‘Delineating potential mechanisms of implicit alcohol cognitions: drinking restraint, negative affect, and their relationship with approach alcohol associations’, Psychology of addictive behaviors: journal of the Society of Psychologists in Addictive Behaviors, Vol. 26, No. 2, pp. 318. Corneille, O. & Stahl, C., 2019, ‘Associative Attitude Learning: A Closer Look at Evidence and How It Relates to Attitude Models’, Personality and Social Psychology Review, Vol. 23, No. 2, pp. 161-189 Davies, M., 2015, ‘Knowledge (Explicit, Implicit and Tacit): Philosophical Aspects’, in International Encyclopedia of the Social & Behavioral Sciences: Second Edition, Volume 13, ed. J.D. Wright, Elsevier. https://doi.org/10.1016/B978-0-08-0970868.63043-X. De Houwer, J., Custers, R. & De Clercq, A., 2006, ‘Do smokers have a negative implicit attitude toward smoking?’, Cognition and Emotion, Vol. 20, No. 8, pp. 12741284.

Implicit Attitude

De Houwer, J.D., 2014, ‘A Propositional Model of Implicit Evaluation’, Social and Personality Psychology Compass, Vol. 8, No. 7, pp. 342-353. https://doi.org/10.1111/spc3.12111. Dehaene, S. & Naccache, L., 2001, ‘ Towards a cognitive neuroscience of consciousness: basic evidence and a workspace framework’, Cognition, Vol. 79, No.1-2, pp.1–37. Degner, J., 2009, ‘On the (un-)controllability of affective priming: Strategic manipulation is feasible but can possibly be prevented’, Cognition and Emotion, Vol. 23, No. 2, pp. 327-354. https://doi.org/10.1080/02699930801993924. Dennett, D. C.,1969, Content and consciousness (2nd ed.), London, Routledge. Dennett, D. C.,1991, ‘Ways of establishing harmony’ in Dretske and his critics ed. B. P. McLaughlin, Cambridge, MA, Blackwell Publishing. Devine, P., 1989, ‘Stereotypes and Prejudice: Their Automatic and Controlled Components’, Journal of personality and social psychology, Vol. 56, No. 1, pp. 5. Devine, P. G., Forscher, P. S., Austin, A. J., & Cox, W. T. L., 2012, ‘Long-term reduction in implicit race bias: A prejudice habit-breaking intervention’, Journal of Experimental Social Psychology, Vol.48, No.6, pp.1267–1278. Di Nucci, E., 2013, Mindlessness, Cambridge, Cambridge Scholars Publishing. Dovidio, J.F. & Gaertner, S.L., 1986, Prejudice, discrimination, and racism, Orlando, Academic Press. Dovidio, J.F. & Gaertner, S.L. ,2004, Aversive Racism, Orlando, Academic Press. Dretske, F., 1969, Seeing and knowing, Chicago, University of Chicago Press. Dretske, F., 1981, Knowledge and the ﬂow of information, Cambridge, MA, MIT Press. Dretske, F.,1988, Explaining behavior, Cambridge, MA, MIT Press Dretske, F., 1990, ‘Does meaning matter?’, Postscript in Information, semantics, and epistemology, ed. E. Villanueva, Cambridge, MA, Blackwell. Dreyfus, H., 2002, ‘Intelligence without representation: Merleau Ponty’s critique of representation’, Phenomenology and the Cognitive Sciences, Vol.1, pp.367–383. Esseily, R., Somogyi, E. & Guellai, B., 2016, ‘The Relative Importance of Language in Guiding Social Preferences Through Development’, Frontiers in psychology, Vol. 7, pp. 1645.

Milena Koot

Evans, J. S. B. T., 2008, ‘Dual-Processing Accounts of Reasoning, Judgment, and Social Cognition’, Annual Review of Psychology, Vol. 59, No. 1, pp. 255-278. https://doi.org/10.1146/annurev.psych.59.103006.093629. Fazio, R. H., 1995, ‘Attitudes as object-evaluation associations: Determinants, consequences, and correlates of attitude accessibility’ in Attitude strength: Antecedents and consequences (Ohio State University series on attitudes and persuasion, Vol. 4), ed. R. Petty & J. Krosnick, Hillsdale, NJ, Erlbaum. Fiske, S. T., 1989, ‘Examining the role of intent: Toward understanding its role in stereotyping and prejudice’ in Unintended Thought, ed. J.S. Uleman & J.A. Bargh, New York, The Guilford Press. Dovidio, J. F., & Gaertner, S. L., 2010, ‘Intergroup bias’ in Handbook of social psychology ed. S.T Fiske, D.T. Gilbert, & G. Lindzey, Hoboken, NJ, John Wiley & Sons Inc. Fiske, S.T., Gilbert, D.T. & Lindzey, G.,2010, Handbook of social psychology, 5th ed., Hoboken, NJ, John Wiley & Sons Inc. Festinger, L., 1957, A theory of cognitive dissonance, Peterson, Evanston, Ill. Figueroa, R. & Harding, S.G., 2003, Science and other cultures: issues in philosophies of science and technology, New York, Routledge. Forscher, P.S., Lai, C.K., Axt, J.R., Ebersole, C.R., Herman, M., Devine, P.G. & Nosek, B.A., 2019, ‘A meta-analysis of procedures to change implicit measures’, Journal of personality and social psychology, Vol. 117, No. 3, pp. 522. Fridland, E., 2014, ‘They’ve lost control: reflections on skill’, Synthese, Vol. 191, No. 12, pp. 2729- 2750. Fridland E., 2015, ‘Learning the way to Intelligence: Reflections on Dennett and Appropriateness’ in Content and Consciousness Revisited, ed. F. De Brigard & C. Munoz- Suarez, Cham, Switzerland, Springer. Fridland, E., 2017a, ‘Automatically minded’, Synthese, Vol. 194, No. 11, pp. 4337-4363. Fridland, E.,2017b, ‘Skill and motor control: intelligence all the way down’, Philosophical Studies, Vol. 174, No. 6, pp. 1539-1560. Fridland, E., 2019, ‘Longer, smaller, faster, stronger: On skills and intelligence’, Philosophical Psychology: Special Issue: Memory and Skill, Vol. 32, No. 5, pp. 759-783.

Implicit Attitude

Gast, A. & Rothermund, K., 2011b, ‘What you see is what will change: Evaluative conditioning effects depend on a focus on valence’, Cognition and Emotion, Vol.25, No.89-110. Gaertner, S.L. & McLaughlin, J.P., 1983, ‘Racial Stereotypes: Associations and Ascriptions of Positive and Negative Characteristics’, Social psychology quarterly, Vol. 46, No. 1, pp.23-30. Gawronski, B., Hofmann, W. & Wilbur, C.J., 2006, ‘ Are “implicit” attitudes unconscious?’, Consciousness and Cognition, Vol.15, No.3, pp.485–499. Gawronski, B. & Payne, B.K., 2010, Handbook of implicit social cognition: measurement, theory, and applications, London, The Guilford Press. Gawronski, B. & Bodenhausen, G.V., 2011, ’The associative-propositional evaluation model. Theory, evidence, and open questions’, Advances in Experimental Social Psychology, Vol. 44, pp. 59-127. https://doi.org/10.1016/B978-0-12-3855220.00002-0. Gawronski, B., & Bodenhausen, G. V., 2014, ‘The associative—Propositional evaluation model: Operating principles and operating conditions of evaluation’ in Dual-process theories of the social mind, ed. J.W. Sherman, B. Gawronski & Y. Trope, New York, The Guilford Press. Gawronski, B., Galdi, S., 2011, ‘Using implicit measures to read the minds of undecided voters’ in Social perception, cognition, and language in honour of Arcuri, ed. M. Cadinu, S. Galdi, A. Maass, Padova, Italy, CLEUP. Gawronski, B., Brannon, S. M., & Bodenhausen, G. V., 2017, ‘The associative-propositional duality in the representation, formation, and expression of attitudes’ in Reﬂective and impulsive determinants of human behaviour, ed. R. Deutsch, B. Gawronski, & W. Hofman, New York, Taylor & Francis Group. Gendler, T.S., 2008, ‘Alief in Action (and Reaction)’, Mind & Language, Vol. 23, No. 5, pp. 552- 585. Glenn, J.J., Werntz, A.J., Slama, S.J.K., Steinman, S.A., Teachman, B.A. & Nock, M.K., 2017, ‘Suicide and self-injury-related implicit cognition: a large-scale examination and replication’, Journal of abnormal psychology, Vol. 126, No. 2, pp. 199. Greenwald, A., Mcghee, D. & Schwartz, J.,1998, ‘Measuring individual differences in implicit cognition: The implicit association test’, Journal of personality and social psychology, Vol. 74, No. 6, pp. 1464-1480.

Milena Koot

Hahn, A., Judd, C.M., Hirsh, H.K. & Blair, I.V., 2014, ‘Awareness of implicit attitudes’, Journal of experimental psychology, Vol. 143, No. 3, pp. 1369. Haith, A. M., & Krakauer, J. W., 2013, ‘Model-based and model-free mechanisms of human motor learning’, Advances in experimental medicine and biology, Vol. 782, pp.1–21. https://doi.org/10.1007/978-1-4614-5465-6_1. Hannan, T.E., Moffitt, R.L., Neumann, D.L. & Kemps, E., 2019, ‘Implicit approach– avoidance associations predict leisure-time exercise independently of explicit exercise motivation’, Sport, Exercise, and Performance Psychology, Vol. 8, No. 2, pp. 210-222. Hehman, E., Flake, J.K. & Calanchini, J., 2018, ‘Disproportionate Use of Lethal Force in Policing Is Associated With Regional Racial Biases of Residents’, Social Psychological and Personality Science, Vol. 9, No. 4, pp. 393-401. Hewstone, M., Rubin, M. & Willis, H.,2002, ‘Intergroup bias’, Annual Review of Psychology, Vol. 53. Heyes, C.M., 2018, Cognitive gadgets: the cultural evolution of thinking, Cambridge, MA, The Belknap Press of Harvard University Press. Holroyd, J. & Sweetman, J., 2016, Implicit Bias and Philosophy, Volume 1, Oxford, Oxford University Press. Hooker, C.A., Penfold, H.B. & Evans, R.J., 1992, ‘Control, Connectionism and Cognition: Towards a New Regulatory Paradigm’, The British Journal for the Philosophy of Science, Vol. 43, No. 4, pp. 517-536. Hurley, S., 2001, ‘Perception And Action: Alternative Views’, Synthese, Vol. 129, No.1, pp. 3-40. Karremans J.C., Stroebe W., & Claus J., 2006, ‘Beyond Vicary’s fantasies: the impact of subliminal priming and brand choice’, Journal of Experimental Social Psychology, Vol.42, No.6, pp.792–98. 62 Kahneman, D., 2011, Thinking, fast and slow, New York, Farrar, Straus and Giroux. Kahneman, D., Slovic, P. & Tversky, A., 1982, Judgment under uncertainty: heuristics and biases, Cambridge, Cambridge University Press. Kahneman, D. & Tversky, A., 1982, ‘Judgments of and by representativeness’ in Judgment under uncertainty: Heuristics and biases ed. D. Kahneman, P. Slovic & A. Tversky, New York, Cambridge University Press.

Implicit Attitude

Kahneman, D. & Frederick, S., 2002, ‘Representativeness revisited: Attribute substitution in intuitive judgment’ in Heuristics and biases: The psychology of intuitive judgment, ed. T. Gilovich, D. Griffin & D. Kahneman, New York, Cambridge University Press Kattner, F., 2012, ‘Revisiting the relation between contingency awareness and attention: Evaluative conditioning relies on a contingency focus’, Cognition and Emotion, Vol.26, pp.166-175. Kawakami, K., Dovidio, J. F., Moll, J., Hermsen, S., & Russin, A. J., 2000, ’Just say no (to stereotyping): Effects of training in the negation of stereotypic associations on stereotype activation’, Journal of Personality and Social Psychology, Vol.78, No.5, pp.871–888. Kawakami, K., Phills, C.E., Steele, J.R. & Dovidio, J.F., 2007, ‘(Close) distance makes the heart grow fonder: Improving implicit racial attitudes and interracial interactions through approach behaviors’, Journal of personality and social psychology, Vol. 92, No. 6, pp. 957. Klauer, K.C., Voss, A., Schmitz, F. & Teige-Mocigemba, S., 2007, ‘Process components of the implicit association test: a diffusion-model analysis.’, Journal of personality and social psychology, Vol. 93, No. 3, pp. 353. Lindgren, K.P., Neighbors, C., Teachman, B.A., Baldwin, S.A., Norris, J., Kaysen, D., Gasser, M.L. & Wiers, R.W., 2016, ‘Implicit alcohol associations, especially drinking identity, predict drinking over time’, Health Psychology, Vol. 35, No. 8, pp. 908. Madva, A., 2016, ‘Why implicit attitudes are (probably) not beliefs’, Synthese, Vol. 193, No. 8, pp. 2659-2684. Mcgrath, A., 2017, ‘Dealing with dissonance: A review of cognitive dissonance reduction’, Social and Personality Psychology Compass, Vol. 11, No. 12. Melnikoff, D.E. & Bargh, J.A., 2018, ‘The Insidious Number Two’, Trends in cognitive sciences, Vol. 22, No. 8, pp. 668-669. Meltzoff, A.N. & Moore, M.K.,1977, ‘Imitation of facial and manual gestures by human neonates’, Science, Vol. 198, No. 4312, pp. 74. Moors, A., 2016, ‘Automaticity: Componential, Causal, and Mechanistic Explanations’, Annual Review of Psychology, Vol. 67, No. 1, pp. 263-287.

Milena Koot

Moors, A. & De Houwer, J., 2006, ‘Automaticity: a theoretical and conceptual analysis’, Psychological bulletin, Vol. 132, No. 2, pp. 297. Moray, N., 1959, ‘ Attention in dichotic listening: Affective cues and the influence of instructions’, Quarterly Journal of Experimental Psychology, Vol.11, pp.56–60. doi:10.1080/17470215908416289. Nagel, J., 2012, ‘Gendler on Alief’, Analysis, Vol. 72, No. 4, pp. 774-788. Ngo, H., 2017, The habits of racism: a phenomenology of racism and racialized embodiment Lanham, MD, Lexington Books. Nock, M.K. & Banaji, M.R., 2007a, ‘Assessment of self-injurious thoughts using a behavioral test’, The American Journal of Psychiatry, Vol. 164, No. 5, pp. 820. Nock, M.K. & Banaji, M.R., 2007b, ‘Prediction of Suicide Ideation and Attempts among Adolescents Using a Brief Performance-Based Test’, Journal of consulting and clinical psychology, Vol. 75, No. 5, pp. 707. Nosek, B.A., Greenwald, A.G. & Banaji, M.R., 2005, ‘Understanding and Using the Implicit Association Test: II. Method Variables and Construct Validity’, Personality & Social Psychology Bulletin, Vol. 31, No. 2, pp. 1661-80. Oswald, F.L., Mitchell, G., Blanton, H., Jaccard, J. & Tetlock, P.E., 2013, ‘Predicting ethnic and racial discrimination: a meta-analysis of IAT criterion studies’, Journal of personality and social psychology, Vol. 105, No. 2, pp. 171. Papineau, D., 2013, ‘In the zone’, Royal Institute of Philosophy Supplement, Vol.73, pp.175–196. Paslakis, G., Kuhn, S., Schaubschlager, A., Schieber, K., Roder, K., Rauh, E. & Erim, Y., 2016,‘ Explicit and implicit approach vs. avoidance tendencies towards high vs. low calorie food cues in patients with anorexia nervosa and healthy controls’, Appetite, Vol.107, pp.171–179. Payne, B. K., & Gawronski, B., 2010, ‘A history of implicit social cognition: Where is it coming from? Where is it now? Where is it going?’ in Handbook of implicit social cognition: Measurement, theory, and applications, ed. B. Gawronski, & B. K. Payne, New York, The Guilford Press. Payne, B.K., Vuletich, H.A. & Lundberg, K.B., 2017, ‘The Bias of Crowds: How Implicit Bias Bridges Personal and Systemic Prejudice’, Psychological Inquiry, Vol. 28, No. 4, pp. 233-248.

Implicit Attitude

Prinz, J., 2004, Gut reactions: A perceptual theory of emotion, Oxford, Oxford University Press. Russell, B., 1912, The problems of philosophy, London, Oxford University Press. Schacter, D., 1987, ‘Implicit Memory: History and Current Status’, Journal of Experimental Psychology: Learning, Memory, and Cognition, Vol. 13, No. 3, pp. 501. Shiffrin, R.M. & Schneider, W.,1977, ‘Controlled and Automatic Human Information Processing: 11. Perceptual Learning, Automatic Attending, and a General Theory’, Psychological review, Vol. 84, No. 2, pp. 127. Shiffrin, R. & Czerwinski, M., 1988, ‘A Model of Automatic Attention Attraction When Mapping Is Partially Consistent’, Journal of Experimental Psychology: Learning, Memory, and Cognition, Vol. 14, No. 3, pp. 562. Spruyt, A., Hermans, D., De Houwer, J., Vandekerckhove, J. & Eelen, P., 2007, ‘On the predictive validity of indirect attitude measures: Prediction of consumer choice behavior on the basis of affective priming in the picture–picture naming task’, Journal of experimental social psychology, Vol. 43, No. 4, pp. 599-610. Squire, L.R., 2009, ‘The Legacy of Patient H.M. for Neuroscience’, Neuron, Vol. 61, No. 1, pp. 6- 9. Stahl, C., Haaf, J. & Corneille, O., 2016, ‘Subliminal evaluative conditioning? Abovechance CS identification may be necessary and insufficient for attitude learning’, Journal of Experimental Psychology: General, Vol.145, pp. 1107-1131. Stanley, J., 2011, Know how, Oxford, Oxford University Press. Stanley, J., & Krakauer, J., 2013, ‘Motor skill depends on knowledge of facts’, Frontiers of Human Neuroscience, Vol.7, pp.503. doi:10.3389/fnhum.2013.0050. Strack, F. & Deutsch, R., 2004, ‘Reflective and Impulsive Determinants of Social Behavior’, Personality and Social Psychology Review, Vol. 8, No. 3, pp. 220-247. https://doi.org/10.1207/s15327957pspr0803_1. Tomasello, M., Carpenter, M., Call, J., Behne, T., & Moll, H., 2005, ‘Understanding and sharing intentions: The origins of cultural cognition’, Behavioral and Brain Sciences, Vol.28, No.5, pp.675–735. Tversky, A. & Kahneman, D., 1974, ‘Judgment under Uncertainty: Heuristics and Biases’, Science, Vol. 185, No. 4157, pp. 1124-1131.

Milena Koot

Van Dessel, P., De Houwer, J., Gast, A., Smith, C. & De Schryver, M., 2016, ‘Instructing implicit processes: when instructions to approach or avoid influence implicit but not explicit evaluation’, Journal of experimental social psychology, Vol. 63. Van Dessel, P., Ye, Y. & De Houwer, J., 2019, ‘Changing deep-rooted implicit evaluation in the blink of an eye : negative verbal information shifts automatic liking of Gandhi’, Social Psychology and Personality Science, Vol. 10, No. 2.

A Logical Challenge to the Future-like-ours Argument and Illegalizing Abortion crystal lam

1 Introduction HIS paper focuses on the Futures-Like-Ours (FLO) argument in the abortion debate. The FLO argument, the idea that â&#x20AC;&#x2DC;because it is wrong to kill beings with futures like ours and since foetuses have futures like ours, therefore abortion is wrongâ&#x20AC;&#x2122;, has occupied much ground of philosophical debate over the ethics of abortion (Cudd 1990; McInerney 1990; Norcross 1990; Sinnott-Armstrong 1999). Don Marquisâ&#x20AC;&#x2122;s influential paper Why Abortion is Immoral (Marquis 1989) first introduces this argument. He argues the loss of a future like ours is sufficient in making the killing of an adult human being wrong and since foetuses also possess such property, therefore the wrongness of abortion can be secured in and is comparable to the wrongness of killing adult human beings. Although that paper only argues for the wrongness and impermissibility of abortion, not concerning whether abortion should be illegalized, it may seem common-sensical to take the FLO as having legal implications: not only suggesting abortion is wrong and impermissible, but also supporting its illegalization. This paper will concern both moral and legal perspectives due to the overwhelming interest in the illegalization and legalization of abortion among politicians, social activists and the general public. This paper hopes to highlight the importance of philosophy in contributing to contemporary political and social debate. This paper will challenge both the position of Marquis and the position from common sense by pointing out three philosophical mistakes: two committed by Marquis and one by common sense. Marquis makes two assumptions in his paper that (I argue) ought to be, but are not, defended. First, he assumes abortion is synonymous to the killing of foetuses. However, the two are not necessarily equivalent. Second, he assumes wrongfulness and impermissibility to be synonymous by using the two interchangeably (Marquis 1989). However, the two are not the same and his paper (if successful) is only able to show the wrongfulness of abortion, not its impermissibility.

British Undergraduate Philosophy Review 1 (2020)

Crystal Lam

The third mistake is committed by common sense, often assuming that if something is impermissible, then it should also be illegal. However, this assumption is questionable in both theory and practice. This paper challenges all three assumptions above. This paper is primarily concerned with the position: if ‘it is wrong to kill beings with futures like ours’, then ‘abortion should be illegal’. To assert that the latter necessarily follows from the former, one must determine what ‘future like ours’ is and answer all following questions affirmatively: Does abortion equate to killing foetuses? Does moral wrongness entail impermissibility? Does moral impermissibility justify enforcing against it? I argue none of these necessarily follow, hence one cannot derive ‘abortion should be illegal’ from ‘it is wrong to kill beings with futures like ours’. Opponents to the FLO argument tend to focus on questioning whether it would be wrong to kill beings with futures like ours (Norcross 1990), whether foetuses do have futures like ours (McInerney 1990) and whether abortion would result in the loss of a future like ours (Sinnott-Armstrong 1999). This paper will show that even if proponents of FLO argument are able to successfully fend off these objections, it would still be in danger. To show this, I grant my opponent the following assumptions: first, I will assume it is wrong to kill beings with futures like ours. The claim is, in fact, sound and intuitive as it makes only a sufficient not necessary requirement for the wrongness of killing: not that it is only wrong to kill beings with futures like ours, but that it is wrong to kill something as long as it has futures like ours (Marquis 1989). Second, I will assume foetuses do have futures like ours. Since foetuses grow (when allowed) into humans, it seems reasonable to assume that they do (or at least will) have futures like ours. If by ‘futures like ours’ we mean futures like that of humans and foetuses are human cells, then it may seem difficult to argue otherwise. Hence, I make two (reasonable) assumptions: first, ‘it is wrong to kill beings with futures like ours’ and second, ‘foetuses do have futures like ours’ (P1, P2 below). It follows from these two assumptions that abortion does result in a loss of a future like ours. This paper will only concern whether it follows ‘abortion should be illegal’, given these assumptions. I will first outline the argument for ‘if it is wrong to kill beings with futures like ours, then abortion should be illegal’, highlighting embedded nuances. Then, I will suggest the premises necessary to derive such a conclusion are not necessarily true, namely: even if it is wrong to kill foetuses, abortion may not be wrong (section 2); even if abortion is wrong, it is not necessarily morally impermissible (section 3); even if abortion is morally impermissible, it should not necessarily be illegal (section 4). Hence, even if one is successful in showing it is wrong to kill beings with futures like ours, one cannot necessarily conclude that abortion should be illegal.

Future-like-ours and Legal Abortion

The argument for if it is wrong to kill beings with futures like ours, then abortion should be illegal is outlined as follows: P1: It is wrong to kill beings with futures like ours P2: Foetuses have futures like ours C1: It is wrong to kill foetuses P3: Abortion is to kill foetuses C2: Abortion is wrong P4: If something is wrong then it is impermissible C3: Abortion is impermissible P5: Impermissible things should be enforced against and made illegal C4: Abortion should be enforced against and made illegal The truth of P1 and P2, thus C1 are granted. This paper will go on to contest P3, P4, and P5, thus undermining C2, C3 and C4. I not only show that the burden of proof is on my opponent to show the premises are necessarily true, but also that to do so would be extremely difficult.

2 Abortion is not necessarily killing foetuses This paper now challenges P3 (abortion is to kill foetuses) to show that even if it is wrong to kill foetuses, abortions may not necessarily be wrong. Let us distinguish between killing and letting die. Killing is directly causing oneâ&#x20AC;&#x2122;s death through active agency. Alternatively, letting die is indirectly causing death through inaction. While both are causally related to death, the degree of agent participation is different: killing directly causes and ensures death through active agency, while letting die indirectly causes death through passive agency. Hence, killing calls for stronger moral condemnation than letting die. Arguably, abortion is closer to letting die than killing. Abortion involves actively aborting, causing the (probable) death of the foetus. Active agency in abortion is the refusal to sustain life by removing the foetus from the womb. Death is an immediate effect of the foetusâ&#x20AC;&#x2122; inability to survive outside the womb, not its removal from the womb, distinguishing it from killing that directly causes and ensures death. This difference is illustrated in Thomsonâ&#x20AC;&#x2122;s violinist scenario (Thomson 1971):

Crystal Lam

You wake up one day connected to a world-renowned violinist with failing kidneys, of which only you can provide. Unplugging him from you will result in his death. Surely, you may unplug yourself from him without committing moral wrongdoing. Choosing to stay plugged-in is morally supererogatory, not obligatory. Despite unplugging yourself from him results in his death, you will not be killing him. You would be denying him life support. What directly causes death is not your action but his health condition. Arguably, abortion is similar to the act of unplugging oneself from the violinist, fundamentally different from killing. Note, it would be a mistake to conclude from above that only abortions that do not result in death are permissible. The key distinction is not the result, but the causal mechanism that brings about the result. Abortion (theoretically) is the removal of the foetus from the womb, which may result in but does not ensure death. On the other hand, to kill is to ensure the death of someone. Killing in the violinist scenario would be shooting him in the head thereby directly ensuring his death, releasing yourself from your relationship. This would be morally impermissible. Chronological sequence is also important. In the original case, you unplug yourself first, which then results in his death. In the killing example, you kill by shooting the violinist first, then unplug yourself from him. Death in the first case is caused by the violinistâ&#x20AC;&#x2122;s own inability to survive without you, while death in the second case is caused by you shooting in him the head. A problematic practical implication may follow from the distinction above between abortion and killing when examining different methods of abortion. In practice, some abortion procedures involve killing the foetus before removing it from the womb. It would, by comparison, be like shooting before unplugging the violinist, hence morally impermissible because death in these cases would be brought about by killing, not refusal of life support as explained above. I accept this practical objection, but I maintain the defence of abortion in the theoretical sense since abortion does not necessarily involve killing foetuses. As long as the method only involves removing the foetus from the womb, not ensuring its death, allowing the possibility of it miraculously surviving outside the womb, then abortion is not necessarily impermissible even if P1 and P2 are true (Thomson 1971).

Future-like-ours and Legal Abortion

3 Wrongfulness does not necessarily entail impermissibility This paper now challenges P4 (if something is wrong then it is impermissible), showing that even if abortion is wrong, it is not necessarily impermissible. Wrong things are not necessarily impermissible. Consider a difficult choice between a wrong and an even worse outcome, one should choose the former over the latter despite both being wrong. Consider sweatshop labour. It is wrongful to allow one to unfairly exploit another’s labour, but would it not be even worse to prohibit those exploited from earning the little income they desperately need? The worse we think it is to unfairly exploit sweatshop workers, the worse we should think of their situation without this option. Faraci (2019) offers such an argument. He argues despite sweatshops being wrong and unjust, they must be allowed to avoid an even worse case in which even the unjustly low sweatshop wage is not available to those who need it most. Hence, something can’t be impermissible simply in virtue of being wrong. Sometimes, it is not only permissible but perhaps obligatory, to choose the wrong over the even worse. This distinction shifts the burden of proof onto the opposing side to show abortion is necessarily the greater evil, thus is morally impermissible. In response, my opponents may meet the burden of proof by highlighting that to allow abortion would violate foetuses’ right to life while to deny abortion would violate mothers’ right to freely dictate what happens in and to her body (I assume there is such right). Since the former is a stronger right than the latter, allowing abortion may seem to be the greater evil, therefore impermissible. However, more analysis is required to determine whether abortion does, in fact, violate foetus’ right to life. Following Thomson (1971), I will suggest it does not, hence is not necessarily the greater evil. Consider Angelina Jolie: I am fatally ill. The only cure is Angelina Jolie’s kiss on my hand. Arguably, Angelina Jolie should kiss my hand because its expected utility far outweighs its cost. Nevertheless, I have no right to her kiss. If she chooses to kiss my hand, that is entirely out of the kindness of her heart. My right to life is at least as strong as the foetus’ and kissing my hand is much less costly than pregnancy. If my right to life does not give me rights to Angelina Jolie’s kiss, then plausibly, foetuses’ right to life also does not give foetuses rights to their mothers’ womb. Different rights have different bearing on the obligation of other parties. Philosophers often distinguish between positive and negative rights. Negative rights are ‘en-

Crystal Lam

title[ments] to non-interference, while positive rights are ‘entitle[ments] to [the] provision of some good or service’ (Wenar 2020). The right against assault is an example of the former and the right to welfare assistance is an example of the latter. Consider Angelina Jolie again, is the right to life in the Angelina Jolie case a positive or negative right? It is necessarily a negative right, the right not to be killed, but whether it is a positive right is not obvious and not yet established. My position is not that beings have no right over anything that sustains life, but only that our right to life does not extend to every life-sustaining thing. Demanding kiss from a stranger and providing life support at costs to oneself are such things that our right to life does not extend over. In causing the death of beings by denying them of such things, their right to life is unviolated, because sustaining their life relies on supererogatory acts. In short, the right to life is necessarily a negative right against being directly/actively killed, not necessarily a positive right to everything that sustains life. Perhaps abortion is wrong, though not because it violates the right to life, but because of the loss of a future like ours. Abortion indirectly causes the death of foetuses by denying life support they have no claim over (just as I do not have a claim to Angelina Jolie’s kiss and the violinist does not have a claim to your kidney). On the other hand, rejecting abortion does violate the right to freely dictate what happens to and in our body (just as Angelina Jolie can decide who to kiss and you can choose to live independently). When given the choice between two wrongs with one violating rights and the other not, there is good reason to choose the latter. Despite the right to life being a stronger right than the right over one’s body, it is not at play in the case of abortion, therefore cannot be cited as a reason for abortion being the greater evil. Hence, even if one accepts abortion to be (passive) killing and thereby wrong, it is not necessarily impermissible. To meet this objection, my opponent not only needs to assert a positive right to life but also a positive right to life that is strong enough to give me a claim to Angelina Jolie’s kiss and the violinist a claim to your kidney. Alternatively, my opponent must suggest other reasons for abortion being the greater evil, thus impermissible. Regardless, the burden of proof is on those making this objection, required to also show that it is impermissible for one to disconnect oneself from the violinist and for Angelina Jolie to refuse to kiss me.

Future-like-ours and Legal Abortion

4 Impermissible things should not necessarily be illegal Finally, this paper challenges P5 (impermissible things should be enforced against), so that even if one grant that abortion is wrong and impermissible, it should not necessarily be enforced against and made illegal. So far, we have been talking about moral judgement; now, P5 leads us into legal waters. The distinction between: 1. Abortion is impermissible and: 2. Abortion should be illegal is needed. Impermissibility concerns morality, illegalization concerns law and enforcement. Unlike moral impermissibility, illegalization necessarily (or should) involve criminalization and enforcement. While some morally impermissible acts such as manslaughter and theft are and should be illegalized, others are not. Impermissibility does not necessarily justify or call for illegalization. While spilling your friendâ&#x20AC;&#x2122;s secrets and cheating on oneâ&#x20AC;&#x2122;s spouse is impermissible, to make these acts illegal would be repugnant. By presenting these cases, I highlight the gap between the impermissibility of an act and whether it should be made illegal. Thus, even if abortion is wrong and impermissible, it is not obvious that it should also be made illegal. Having shown cases of morally impermissible things that should not be made illegal, the burden of proof is again on the opposing camp to show abortion necessarily falls outside this category. An objection can be raised by suggesting that if one opposes illegalizing abortion despite P1, P2, P3, P4 being true, it is because one is not confident enough in the truth of the assumptions P1, P2, P3, P4. If we know and genuinely believe that something is wrong and impermissible, for sure, then what reasons are there against prohibiting it (Devlin 1959)? Why let people get away with such impermissible wrongdoing without any (guaranteed) repercussion? Perhaps there is truth to this challenge, as even when granting the truth of P1, P2, P3, and P4, I am not genuinely convinced as, I trust, many others are not. But, even if one assumes P1, P2, P3, and P4 to be true without a doubt, is there no reason against illegalizing abortion? I suggest there is. Not every impermissible act is and should be illegal. For instance, deception between individuals, not concerning the society at large, such as cheating on oneâ&#x20AC;&#x2122;s spouse, is and should not be criminalized. However, deception concerning social order such

Crystal Lam

as plagiarism and breaching of contracts are and should be criminalized. It seems there are things only warranted to first-party participants, not third-party observers/legislators, making first and third-party rights different. Perhaps only when matters concern third parties (society) should third parties (society) be given rights to enforce it (Hart 2002). In addition to highlighting impermissibility does not necessarily entail and call for illegalization thus shifting the burden of proof back onto my opponent, I suggest a positive reason for not illegalizing all impermissible acts. The reason lies behind the notion of a perfect society: perfect societies are safe, stable and respectful due to the sheer moral will of its people not the stringency of its laws. When we educate our children, we not only want them to be dutiful but more importantly, good. Perhaps this explains our reluctance to prohibit all morally impermissible acts: our desire to give people the largest degree of freedom possible (given social order is unharmed) so that they do good out of their desire for good, not fear of punishment. Hence, even if P1, P2, P3, and P4 are true, one can still reasonably reject illegalizing abortion.

5 Conclusion Motivated by assumptions made by Marquis (1989) and common sense, this paper set out to answer the question ‘if it is wrong to kill beings with futures like ours, should abortion be illegal?’. While to answer affirmatively may seem natural as first glance, this paper argues the contrary by identifying and analysing the necessary logical steps between accepting the claim that ‘it is wrong to kill beings with futures like ours’ and concluding ‘abortion should be illegal’. To derive the latter from the former, one needs to first argue abortion is wrongful by showing that foetuses are beings with futures like our, second equate abortion with the killing of foetuses. Following from this, to argue abortion is impermissible, one needs to either argue wrong things are impermissible or show that even though not all wrong things are impermissible, abortion is such thing that is both wrongful and impermissible. Finally, to argue abortion should be illegal, one must either show that all impermissible things should be made illegal or suggest even though not all impermissible things should be made illegal, abortion is such thing that is impermissible and should be illegal. This paper tracks every step of the logic above, granting the assumption that foetuses are beings with futures like ours and (most) abortion results in the loss of a future like ours, I argue none of the other logical steps necessarily follows: abortion is not necessarily the killing of foetuses, thus abortion is not necessarily wrong; wrong things are not necessarily impermissible, thus even if abortion is wrong, it is not necessarily impermissible; impermissible things

Future-like-ours and Legal Abortion

should not necessarily be enforced against, thus even if abortion is impermissible, it should not necessarily be enforced against. I provide reasons at each stage of reasoning so that even if the previous layer of defence, fails, another awaits. To conclude, even if it is wrong to kill foetuses (that have futures like ours), abortion should not necessarily be made illegal. University College, Durham, England

References Cudd, A. E. (1990). Sensationalized philosophy: A reply to Marquis’s ‘Why Abortion is Immoral’. Journal of Philosophy 87, 262–264. Devlin, P. (1959). The enforcement of morals. Proceedings of the British Academy. Neither the Author nor Typesetting Editor were able to find further bibliographic details. Faraci, D. (2019). Wage exploitation and the nonworseness claim: Allowing the wrong, to do more good. Business Ethics Quarterly 29(2), 169–188. Hart, H. L. A. (2002). Law, Liberty and Morality. Stanford. Marquis, D. (1989). Why abortion is immoral. Journal of Philosophy 86(4), 183–202. McInerney, P. K. (1990). Does a fetus already have a future-like-ours? Journal of Philosophy 87(5), 264–268. Norcross, A. (1990). Killing, abortion, and contraception: A reply to Marquis. Journal of Philosophy 87(5), 268–277. Sinnott-Armstrong, W. (1999). You can’t lose what you ain’t never had: A reply to Marquis on abortion. Philosophical Studies 96(1), 59–72. Thomson, J. J. (1971). A defense of abortion. Philosophy and Public Affairs 1(1), 47–66. Wenar, L. (2020). Rights. In E. N. Zalta (Ed.), The Stanford Encyclopedia of Philosophy (Spring 2020 ed.). Metaphysics Research Lab, Stanford University.

Hermeneutical Injustice, Locutionary Silencing, and Bad Sex lorenz von lüpke

1 Introduction REVIOUS philosophical work on sexual consent has mainly focused on the extreme or very bad cases, like rape or coerced sex, where someone’s dissent is ignored or failed to be taken as such. Relatedly, sexual consent has been presented as a binary choice between an explicit yes and an explicit no. However, in this essay I present two cases of bad sex where explicit consent is absent, but that also do not fit with cases of explicit dissent found in the consent literature. In section 2, I briefly argue that the consent in these cases is invalid because it is not the result of autonomous choice and is harmful to agency, and in section 4 I explain why, given that the consent is invalid, the agents did not explicitly dissent. I suggest that this can be explained referencing the work of Alessandra Tanesini on locutionary silencing and of Miranda Fricker on hermeneutical injustice, the injustice of lacking the shared conceptual resources to make sense of one’s experience. I thereby show how hermeneutical injustice and locutionary silencing are connected: together they render an agent both unable to understand her experience and wants and unable to articulate these, which explains why the agents in my cases do not explicitly dissent. Further, the structures of social oppression that produce these cases also impede possibilities of overcoming these dynamics. These findings lead me to conclude that theories of sexual consent must take into account the ways in which structures of social oppression can influence whether and how sexual consent is valid. A brief terminological note: I use the word ‘women’ to pick out anyone who has been socialised to be a woman, and anyone who identifies themselves with the category ‘woman’.

2 Two cases Consider these cases: British Undergraduate Philosophy Review 1 (2020)

Silencing and Bad Sex 1) Sarah and Fred have been in a relationship for two years and genuinely care for each other. One evening, Sarah is tired from work and wants to watch her favourite TV show Downtown Abbey, but Fred makes sexual advances. She is not in the mood for sex. But because she knows that he is prone to foul moods when he does not get what he wants and because she does not want to have to deal with that, she concedes to Fred’s advances. 2) Mara and Linda are two teenagers and have been dating for three months. Neither has had sex before and Linda tells Mara that she would like to try it with her. Mara is unsure whether she would like to have sex yet but feels obligated because she is afraid of disappointing Linda, and because she thinks that sex is necessary for having a real relationship. Although she would prefer it if nothing sexual happened between them yet, she steels herself and initiates the sex to get it over with.

In both cases, something seems wrong with the sexual consent of Sarah and Mara, though they do not explicitly dissent. Neither case involves overt pressuring or coercion, but explicit sexual consent is nevertheless not given either, at least none that expresses that Sarah and Mara also want to do what their partners want. Their consent resembles more a concession or compromise than an expression of want. I argue that that is not enough to qualify as valid sexual consent. This is despite the fact that they at least consent in speech, and arguably even in thought. However, their consent in thought seems troubling as it is based on the wrong reasons and does not stem from a real desire to engage in the acts. The choice to engage in the sexual acts is not autonomous. Further, sexual consent of that kind is insufficient to avoid harm to the participants resulting from the sexual act. Indeed, plausibly, harm to a person’s agency can arise precisely from the experience of conceding to undesired sex since that kind of concession involves relinquishing control over one’s bodily autonomy, exactly because the acts done to and involving the body are not truly desired. The harm resulting from these sorts of sexual experiences and specifically the harm to a person’s agency leads me to think that the consent in the cases described is invalid, even if the agents consent in speech and thought. But why are Sarah and Mara not voicing their dissent? In the following, I draw on Tanesini’s account of silencing in connection with Fricker’s account of hermeneutical injustice to explain what is going wrong in these cases, focusing on the failure to communicate dissent. I argue that these are cases of locutionary silencing (the literal prevention from speaking), where the silencing stems not primarily from the agent’s conversation partner’s actions but from structural, societal dynamics that

Lorenz von Lüpke

silence the agent. In the next section, I explain first locutionary silencing and its relationship with timidity and servility, and then hermeneutical injustice.

3 Locutionary silencing and hermeneutical injustice Alessandra Tanesini in her 2016 article ‘Calm down, dear: intellectual arrogance, silencing and ignorance’ gives an account of locutionary silencing and its relationship with the intellectual vices of timidity and servility. She traces the development of timidity and servility back to repeated exposure to a conversation partner’s intellectual arrogance, claiming that the recipient of that behaviour is more likely to develop timidity and/or servility (Tanesini 2016: 86). One of Tanesini’s examples of arrogant behaviour is the disregard for rules of turn-taking in conversation. As an individual repeatedly experiences being locutionarily silenced through interruption or being denied a chance to speak, she is more likely to become intellectually timid (Tanesini 2016: 76). An agent manifests the timidity if she is resigned to receiving less respectful treatment and is afraid of defending her position, tends to hide to avoid being noticed or being expected to contribute to the conversation (Tanesini 2016: 88). Servility, on the other hand, is characterizad as an ‘excessive deference to the views held by other individuals’ and manifested by mistrusting one’s own judgement and instead following other people’s views (Tanesini 2016: 89). The two vices silence the agent who manifests them as they prevent her from speaking. Although they are agent-internal constraints, Tanesini shows how they actually result from external pressures. These pressures do not have to exist in every interaction. She writes that the silencing can come from adaptive preferences, which are preferences that stem not from autonomous choice but from circumstances of oppression, which the agent accepts and adapts her preferences to.(Tanesini 2016: 87) This reveals the link to structural domination and oppression. Fricker’s hermeneutical injustice operates exactly on the level of structural oppression, such that there is no one person committing the injustice (Fricker 2007: 159). Hermeneutical injustice is a type of epistemic injustice that occurs when two conditions are met: 1) the collective pool of concepts lacks a concept that would enable a speaker to make sense of a significant part of her social experience, and 2) this lacuna results from the systematic marginalization of the speaker’s social group in the process of hermeneutical production such that they do not have sufficient opportunity to influence the process so as to create concepts reflecting their experience (Fricker 2007: 155). Hermeneutical production is the process of creating shared conceptual resources

Silencing and Bad Sex

used by members of the society to interpret and make sense of their experiences. Researchers, lawmakers, journalists, advertising specialists, and others are involved in that process. Fricker suggests that hermeneutical injustice can be so damaging that it hinders the construction of selfhood and suggests that this in part constitutes the primary harm of hermeneutical injustice (Fricker 2007: 168). In the context of sexuality and consent, hermeneutical injustice might hinder the development of one’s sexuality, of an awareness of one’s wishes and abilities to communicate them. As an illustrative example, until 1991 marital rape was exempt from legal prosecution in England and Wales as spouses were considered to possess a right to sex with their partner. Clearly, the idea of full bodily and sexual autonomy for women is there not incorporated into their sexual identity. As lawmakers are important figures in the process of hermeneutical production, this provides an example of how women’s sexual experience was (and is) hermeneutically marginalizad.

4 Locutionary silencing, hermeneutical injustice, and bad sex In this section, I show how Tanesini’s work on locutionary silencing and Fricker’s on hermeneutical injustice are connected and how they can explain why Sarah and Mara did not explicitly dissent in the two cases above. As Tanesini clearly states, the phenomenon of arrogance leading to the development of timidity and/or servility and thereby to locutionary silencing is not due to vicious individual psychologies but ultimately comes from systemic ‘social relations of domination and subordination’ (Tanesini 2016: 86). The relations between men and women are such oppressive relations. From a social constructivist view of gender, social norms are constitutive of gender categories, since it is not biological differences between male and female humans giving rise to different behaviour, but instead the process of gendered socialization creating these differences. Gender is constituted by social norms. In a patriarchal society such as ours, the social norms constituting the category ‘woman’ subordinate the people socializad to be women. As Tanesini shows, people in positions of dominance (i.e. men in this context) exhibit behaviours that condition those in subordinated positions into silence. Therefore, in a patriarchal society, the norms constituting the category ‘woman’ silence. Even without Fred or Linda coercing or threatening them, previous experiences and socialisation likely made Sarah and Mara less confident in speaking up for themselves and their wants. The norms regarding communication are important, but it is the further norms regarding women and sexuality that create the specific context for these cases. These

Lorenz von Lüpke

also fit into the pattern of subordination. Sexual norms for women include ideas of constant availability, of prioritising their (male) partner’s pleasure over their own, of being an object of someone else’s sexual lust and desire without the agency to direct or control that lust. The lack of adequate concepts for women’s sexuality make it harder for women to express their wants and deny what they do not want. For centuries, women have not had the hermeneutical resources to make sense of their experiences. Concepts such as the female orgasm, female sexual pleasure, and female agency in sex are still new or contested, the #MeToo movement illustrates that these ideas of female sexuality as constantly available and as a service are still all too prevalent, while a woman’s sexual consent is secondary. Society at large is grappling with the absence of adequate concepts to describe the sexual experience of many women. Likely, neither Sarah nor Mara is fully able to articulate to herself what is happening in these situations, beyond a diffuse sense that something is amiss. In conjunction with the communicative norms that make women less likely to speak up for themselves, they are facing two obstacles they would need to overcome in order to express their dissent in the cases described. The sexual norms for women and absence of adequate concepts, in conjunction with the communicative norms, create situations where women are incapable of expressing their wishes, and thus they are locutionarily silenced. The cases I described show how there is a connection between norms of communication that lead to locutionary silencing on the one hand and the hermeneutical injustice as the absence of concepts on the other hand. When (1) the concepts to describe and communicate one’s experience are not readily available, and (2) the agent through dominating and coercive experiences in communication has acquired the intellectual vice(s) of timidity and/or servility, the outcome is a situation where the individual is unable to speak of her experience and voice her wants. She is locutionarily silenced and plausibly also less likely to comprehend what it is she experiences, both (a) in regard to what she would like to communicate, and (b) in regard to the experience of being silenced. If the concepts of silencing and hermeneutical injustice concerning women’s sexuality became part of the collective pool of shared resources, Mara and Sarah could understand (b), which likely would help them in understanding and communicating (a) as well. The obstacle of the communicative norms that silence them in general would not be removed, however, and neither would the absence of adequate concepts for women’s sexuality. Indeed, Fricker describes how the structures of oppression that create hermeneutical injustice also work to prevent overcoming that injustice. She points to a link between hermeneutical injustice and what she terms testimonial injustice. Testimonial

Silencing and Bad Sex

injustice occurs when a speaker is given less credibility by a hearer than she deserves, owing to a prejudice about her social group on the part of the hearer (Fricker 2007: 28). Fricker’s concept of testimonial injustice describes an additional way in which women are dizadvantaged in the sphere of communication and exchange of information and connects well with Tanesini’s account of silencing communicative norms. Tanesini focuses on the effects of intellectual arrogance in terms of the development of intellectual vices on the part of the hearer. Fricker’s testimonial injustice can also be conceived of as capturing the effect of intellectual arrogance, but instead of describing the intellectual vices that develop from it, she focuses on the moral and epistemic injustice that results from it. What Fricker finds in the connection between testimonial and hermeneutical injustice is that the two injustices can combine to make it even harder for the speaker to be heard and assigned the correct amount of credibility. Credibility for someone’s testimony on a topic where there are inadequate hermeneutical resources will already be lower because of the hermeneutical injustice. In combination with a prejudice against her social group on the part of the hearer, the attempt to articulate a badly understood experience will be assigned even lower credibility, which reduces the chances that the hermeneutical injustice will be alleviated (Fricker 2007: 159). Tanesini’s focus on the development of intellectual vices enables her to trace the long-term effects of exposure to intellectual arrogance, which Fricker’s testimonial injustice does not. She describes how the effects of exposure to intellectual arrogance can become entrenched into someone’s character. This can further increase the effect of the double injustices Fricker describes. Through repeated exposure to intellectual arrogance, a speaker becomes servile and/or timid. She has an experience that she lacks the concepts to describe adequately. She has had sufficient experiences with attempting to speak even on topics where she had adequate conceptual resources but experienced intellectual arrogance and is so less likely to speak up for herself and her experience. This means she is less likely to even raise the topic of that for which adequate hermeneutical resources are lacking. If she were to raise it, Fricker’s testimonial injustice compounded with the hermeneutical injustice mean that it is less likely that she would be given adequate credibility for her experience. But Tanesini’s account of silencing explains how it is that the experiences do not become raised at all. The individuals – like Sarah and Mara in the cases above – censor themselves. And this is why it so difficult to overcome the hermeneutical injustice under circumstances of oppression that also shape how members of an oppressed group are treated in the context of communication and knowledge sharing.

Lorenz von Lüpke

One could ask whether the oppressive norms regarding communication and sexuality possess as strong of a grip on women as I have indicated in this essay. Is it not possible for women to reject these norms and cease to be subordinated in these ways? After all, these are only norms, and norms can be rejected. However, the problem is that these social norms are constitutive and not merely regulative of the gender categories. This means that these social norms are what it means to be a woman. It is not only that society wants women to conform to these norms and punishes those who do not. It means that to be a woman is to adhere to these norms. What a woman is in a society like ours is described by the norms. And to be socializad into being a woman means internalising them. The experiences of people who consciously denounce these norms and find themselves re-enacting them regardless support this. Even people who cease to identify with the category ‘woman’ can find themselves enacting these norms because of the socialization they received. That is because the self is constituted by socialization over a long period of time, and socialization is gendered. If the norms were merely regulative, individuals would be able to simply reject them and have nothing to fear besides criticism. But the norms are constitutive of the gender categories and to reject them is far more complicated than simple rule breaking and subsequent punishment. It involves not being recognizad as a woman anymore. Since the norms are internalizad and the self is gendered, it is difficult to let go of them. Further, the purpose of making both agents in my second example case women was to show that these norms do not only exist in the relation between men and women, but also exist in the absence of a male agent. These norms are constitutive of what it means to be socializad to be a woman in our society: lacking the concepts to be able to articulate their sexual experience, and silenced through the intellectual vices of timidity and servility.

5 Conclusion In this essay I have argued that valid consent is missing in the two cases I described and shown that there is a connection between Fricker’s hermeneutical injustice and Tanesini’s account of locutionary silencing. The connection is such that women as agents in oppressive structures are not in possession of adequate concepts to describe their sexual experiences and wants and are socializad as women to develop timidity and servility and be thereby locutionarily silenced. Together, these dynamics explain how and why valid sexual consent is missing in the cases I described, where women acquiesce to sexual acts they do not desire. These findings suggest to me that theories of sexual consent must take into account the ways in which structures of social oppres-

Silencing and Bad Sex

sion can influence whether and how sexual consent is valid, otherwise they cannot explain why sexual consent is invalid in the cases I described. The University, Glasgow, Scotland

References Fricker, M. (2007). Epistemic Injustice: Power and the Ethics of Knowing. Oxford. Tanesini, A. (2016). “Calm down, dear”: Intellectual arrogance, silencing and ignorance. Aristotelian Society Supplementary Volume (1), 71–92.

Testimonial Injustice in Virtue of Due Credibility viola nassi

1 Introduction HE concept of epistemic injustice has recently arisen as a major topic in the field of epistemology, connected to testimony and credibility around discussions of social power dynamics (Grasswick 2018). In Epistemic Injustice: Power and the Ethics of Knowing (2007), Miranda Fricker presented the notion of testimonial injustice (henceforth, T.I.). Roughly, T.I. is an epistemic wrong that occurs when a hearer is not afforded enough credibility by a speaker as a result of an identity prejudice, where the hearer suffers what Fricker dubs a credibility deficit. However, Federico Luzzi (2016) suggested that there could be cases of T.I. where the hearer is given the right level of credibility, and thus that the original definition of T.I. may prove too narrow. In this paper, I will advance Luzzi’s proposal. In fact, I contend that there are cases where T.I. obtains in virtue of due credibility. That is to say, I wish to introduce a novel kind of T.I., to suggest that we may have to consider a definition of T.I. that does not require a credibility deficit. In Section 2, I will introduce Fricker’s view of T.I. In Section 3, I will refer to Luzzi’s proposal to observe a particular case of mansplaining which I believe captures a novel type of T.I. where the speaker is epistemically harmed in virtue of being afforded the right amount of credibility. In the same section, I will confute a possible objection to my argument, to conclude in Section 4 that we could plausibly embrace a broader definition of T.I.

2 Miranda Fricker’s view I will now introduce Miranda Fricker’s account of T.I. For Fricker, T.I. involves ‘a prejudice through which the speaker is misjudged and perceived as epistemically lesser’ (Fricker 2017: 53). Accordingly, central cases of T.I. encompass an identity prejudice. The latter consists in a kind of prejudice related to one’s social identity, which tracks British Undergraduate Philosophy Review 1 (2020)

Testimonial Injustice and Due Credibility

people in relation to other types of injustice, such as racial and gender injustice. When an identity prejudice brings about T.I., T.I. is then systematically linked to other kinds of injustice (Fricker 2007: 27). The peculiarity of T.I. is that it harms the speaker not only ethically, but also epistemically, as she is ‘wronged speciﬁcally in her capacity as a knower’ (Fricker 2007: 20). For instance, consider a passer-by asking a woman for information about where a certain location is, to then ask the same question to a man nearby. Imagine that the passerby chooses to believe the man rather than the woman perhaps because, according to negative stereotypes about women, he falsely thinks that her sense of direction is poor. Hence, the passer-by has no evidence as to why he should believe the man rather than the woman, as an identity prejudice leads him to accept the man’s testimony. As a result, the passer-by discredits the woman’s knowledge, wronging her as an epistemic agent. In T.I., identity prejudice distorts the hearer’s credibility judgment regarding the speaker’s testimony (Fricker 2007: 36). Credibility plays a fundamental role in cases of T.I.: as Fricker illustrates, in testimonial exchanges ‘the hearer’s obligation is obvious: she must match the level of credibility she attributes to her interlocutor to the evidence that he is offering the truth’ (Fricker 2007: 19). This is exactly what goes amiss in instances of T.I., as Fricker explains that identity prejudice can either deflate or inflate the credibility given to a speaker. Fricker considers two central types of prejudicial dysfunction happening in T.I.: credibility deficit and credibility excess (Fricker 2007: 17). In the first case, the speaker will be afforded less credibility than she is due, whilst in the second she will be given too much credibility. Fricker concentrates on credibility deficit as she deems that, unless concerning specific cumulative cases, credibility excess tends to be advantageous (Fricker 2007: 19), whilst credibility deficit wrongs the speaker each and every time it occurs, being harmful even in token cases. It is now relevant to unpack Fricker’s notion of credibility, as I will later return to it when evaluating Fricker’s view in Section 3. Fricker contends that credibility judgments are a matter of epistemic trustworthiness, which in turn requires the conjunction of two components: competence and sincerity (Fricker 2007: 45). When making a credibility judgment, a hearer assesses whether she takes the speaker to be competent, and whether she takes him to be sincere. In cases of T.I., the hearer’s identity prejudice can affect one or both components of trustworthiness, in either case wronging the speaker in her capacity as a giver of knowledge. Finally, we obtain a definition of Fricker’s account of T.I. as follows: T.I. is a kind of epistemic injustice that occurs ‘if and only if [the speaker] receives a credibility deficit owing to identity prejudice in

Viola Nassi

the hearer; so the central case of testimonial injustice is identity-prejudicial credibility deﬁcit’ (Fricker 2007: 28).

3 Testimonial injustice in virtue of due credibility In his paper Testimonial Injustice Without Credibility Deﬁcit (or Excess), Luzzi briefly mentions a parallel between one of the cases he devises and an instance of what is called mansplaining (Luzzi 2016: 208). In this section, I will present a specific type of mansplaining to show that it illustrates what I will henceforth call T.I. in virtue of due credibility. To begin with this discussion, it will be useful to give a dictionary definition of mansplaining itself. In the OED, mansplaining is defined as follows: ‘to explain (something) needlessly, overbearingly, or condescendingly, esp. (typically when addressing a woman) in a manner thought to reveal a patronizing or chauvinistic attitude’ (OED: ‘mansplain’). I will now delve into a feasible case of mansplaining, to then present my reasons for arguing that in this case T.I. happens in virtue of due credibility. Imagine an engineer in conversation with a man. While she is explaining the project she is working on, he repeatedly interrupts her by continuing to say exactly what she was stating, manifesting an intention to overpower her. The man then only repeats what the engineer was saying with his voice, adding nothing to the conversation in place. To understand better how such an exchange constitutes an instance of mansplaining, let us underline that the man’s ‘explaining’ consists of re-elaborating what the engineer is asserting, believing that he is explaining it correctly and in a better way. Therefore, although the man’s main purpose and conviction might be that of providing an explanation, the result is not that of an enriched exchange, due to the man’s intention to overpower his interlocutor. What is obtained thus, can be called a ‘mansplanation’. Moreover, say that the man is a colleague of the woman’s, but their expertise diverges. It follows that he is aware of his interlocutor’s profession and epistemic status, though he nonetheless, perhaps even incorrectly, ‘mansplains’ her own project to her. The engineer suffers a T.I.: she is wronged epistemically on the basis of an identity prejudice, the harm is systematic, and it is not strictly cumulative: this is a one-off situation which can and will present itself again, for these cases happen routinely. The wrong she suffers also involves the ‘discriminatory but ingenuous misjudgement’ Fricker considers relevant in cases of T.I. (Fricker 2017: 54). The prejudicial dysfunction here results in the speaker being given the right amount of credibility (Luzzi 2016: 204): no credibility deficit is present. Furthermore, it is pre-

100

Testimonial Injustice and Due Credibility

cisely in virtue of the right amount of credibility being given that the engineer suffers a T.I. In fact, in our case, the hearer wrongs the speaker on account of his appropriation of her credibility. If the hearer did not see the speaker as trustworthy, the epistemic harm would actually not obtain. That is, if our engineer was not seen as sincere and competent in making her claims, her hearer would not re-assert them, and he would thus not afford her the credibility she is due. Therefore, the engineer’s interlocutor re-appropriates her credibility and role as a credible knower. As a matter of fact, if he did not believe what she said, he would not take ownership of her very claims via re-asserting them to begin with. It is clear that credibility is not affected in any way in this case. Hence, in the instance we have examined, T.I. obtains not despite the hearer being afforded due credibility, but exactly in virtue of it. Our engineer is harmed as a giver of knowledge exactly because her interlocutor fully believes she is credible. Mansplainers act this way because of their presumed authority over women. Said authority exists simply on account of their gender, as mansplainers manifest an identity prejudice by viewing women as innately epistemically lesser. In Men Explain Things to Me, the essay that inspired the coinage of the term mansplaining, Rebecca Solnit describes a male interlocutor as having ‘that smug look I know so well in a man holding forth, eyes fixed on the fuzzy far horizon of his own authority’ (Solnit 2014: 2). This deluded gender-based authority generates epistemic hubris, which is inflated negative epistemic privilege1 (Baird and Calvard 2019: 270). Epistemic hubris is closely connected to power, arrogance and unjustified over-confidence concerning knowing (Baird and Calvard 2019: 270–271). Accordingly, the mansplainer’s self-assumed authority is what leads him to regard the woman as a lesser knower, whilst still allowing him to credit her for her claims. The mansplainer accepts the woman’s testimony, but he does not regard her as authoritative, and this is precisely what comes about in our mansplaining case of T.I. I will now present a possible counter-response to my argument to maintain the original definition of T.I. as requiring a credibility deficit. A defender of the original notion of T.I. could respond that it is not clear whether the speaker in the case I propose is treated to be credible at all. After all, let us recall that Fricker requires that the hearer’s obligation in testimonial exchange is to ‘match the level of credibility [he] attributes to [his] interlocutor to the evidence that [she] is offering the truth’ 1 Steven Ogden first introduced this notion in The Church, Authority, and Foucault (2017), attributing it to collectives and specifying that epistemic hubris is ‘not about personal pride. Instead, it is an exaggerated sense of epistemic privilege, which is symptomatic of sovereign power’ (Ogden 2017: 16). However, this concept can be employed herein discussing individual epistemic privilege, which is still, as implied by the concept of identity prejudice, a product of dominant gender power.

Viola Nassi

101

(Fricker 2007: 19). Possibly, the evidence that the engineer is offering the truth to the mansplainer could not match the credibility he gives her. One could suggest that the reason why the engineer’s interlocutor mansplains is that he does not take her to be sincere. Perhaps he thinks women are prone to lie more, and he feels the need to reassert her claims, as with his mansplanation, he erroneously thinks that he is explaining them in the correct way. In this case, he would violate one of the two components of trustworthiness as discussed above. This would be enough on Fricker’s account to argue that the engineer actually suffers a credibility deficit (Fricker 2007: 44-45), as she would be discredited first, to only then be robbed of her credibility. Hence, the case I introduce would not be counted as an instance where due credibility is given to begin with. Nonetheless, this objection does not quite succeed. It is improbable to imagine that the engineer is not seen as trustworthy by her interlocutor. Let us recall that Fricker embraces the idea that credibility judgments are matters of perception (Fricker 2007: 66), where the latter is influenced by one’s pre-existing beliefs. The hearer sees the speaker as either trustworthy or not. In the case of mansplaining I examined, women are actually perceived to be sincere, and it is not the assessment of their competence that is affected, but their presumably inferior authority and status on account of their gender. Therefore, the claim that trustworthiness, and thus credibility judgment, is not affected still stands: mansplainers of the kind we examined epistemically harm women in virtue of the credibility they afford them. It appears reasonable to state that the original account of T.I. as requiring a credibility deficit should be broadened to include cases of T.I. which obtain in virtue of due credibility.

4 Conclusions In conclusion, in this paper I have further explored Luzzi’s suggestion that T.I. cases with due credibility are not only possible, but likely to exist. Moreover, I have shown that it is plausible to imagine T.I. cases that obtain precisely in virtue of due credibility being afforded to the speaker. It proves then that we have solid reasons to hold that Fricker’s original conception of T.I. may be too narrow. As a result, we have grounds for considering a broader conception of T.I. which incorporates this novel kind of T.I. in virtue of due credibility. The University, Glasgow, Scotland

102

Testimonial Injustice and Due Credibility

References Baird, C. and T. S. Calvard (2019). Epistemic vices in organizations: Knowledge, truth, and unethical conduct. Journal of Business Ethics 160(1), 263–276. Fricker, M. (2007). Epistemic Injustice: Power and the Ethics of Knowing. Oxford. Fricker, M. (2017). Evolving concepts of epistemic injustice. In I. J. Kidd, J. Medina, and G. P. Jr. (Eds.), Routledge Handbook of Epistemic Injustice, pp. 53–60. Routledge. Grasswick, H. (2018). Feminist social epistemology. In E. N. Zalta (Ed.), The Stanford Encyclopedia of Philosophy (Fall 2018 ed.). Metaphysics Research Lab, Stanford University. Luzzi, F. (2016). Testimonial injustice without credibility deficit. Thought: A Journal of Philosophy 5(3), 203–211. Ogden, S. G. (2017). The Church, Authority, and Foucault: Imagining the Church as an Open Space of Freedom. Taylor & Francis. Solnit, R. (2014). Men Explain Things to Me, pp. 1–15. Haymarket.

A Case Against Dreaming as Hallucination willa saadat

ALLUCINATIONS entail an agent having false perceptual experiences— having sensations of phenomenon that are not actually occurring. It therefore seems natural to believe that dreaming is a kind of hallucinatory experience. However, I will argue that this approach is mistaken, drawing on conceptual analysis of dreams as well as empirical studies and literature. This essay will explore and defend McGinn’s and Ichikawa’s account of dreams as imaginative experiences and use arguments from empirical data conducted by dream scientists like David Foulkes to suggest a key problem with the hallucinatory model of dreaming—that it is undermined by studies reporting connections between imaginative ability and dream experiences whilst the imagination model is able to accommodate this evidence. In order to show that dreams are not hallucinations, I will argue that dreams do not involve percepts and instead, as Foulkes suggests, dreams are a form of complex, intelligent behaviour with cognitive prerequisites. From this, I will argue that the imagination model of dreams fits and is supported by Foulkes’ empirical findings. This will enable me to conclude that we have strong reason to reject the view that dreams are hallucinations and adopt the view that they are imaginative experiences. There is an intuitive appeal to the view that dreams are hallucinations—wakeful hallucinations and dreams seem to share very similar (if not identical) phenomenology. Both involve an agent seeming to have perceptions or other kinds of sensory experiences that do not correspond to the external world, thereby suggesting that dreaming involves percepts (the kind of sensory experiences that one has when they are wakefully conscious) (Ichikawa 2009: 104). As Hobson puts it, there is no difference in the nature of his waking experience and his dream experience besides the obvious fact that the former corresponds to reality and the latter does not—in his dream he merely had hallucinations, seeing and hearing things that were not in his bedroom (Hobson 1999: 5).

British Undergraduate Philosophy Review 1 (2020)

103

Against Dreaming as Hallucination

104

Now, since the percepts an agent has in their dream are misleading and they are having sense experiences of phenomena that are not actually occurring, we can understand these false percepts to be hallucinations. After all, if hallucinations are characterized as experiences that are subjectively indistinguishable from veridical sense perceptions but without the agent genuinely having sense perception of the phenomenon, then this is what seems to be occurring in dreams. For someone to hallucinate x, this would entail that they have a perceptual experience of x despite x not being the case or genuinely occurring. Likewise, for someone to dream y, this would entail they have a perceptual experience of y despite y not actually occurring. According to the hallucination model, y is an instance of x. So, the argument for the view that dreams are hallucinations can be characterized as: 1. Dreaming and waking perception both involve percepts. 2. In our dreams, our percepts are misleading: we have sense experiences of phenomena that are not actually occurring. 3. To have false or misleading percepts is to have a hallucination. 4. Therefore, to dream is to, at least in part, hallucinate. Premises (1), (2) and (3) are not without contention and those wishing to challenge the hallucination model can deny each, or all, of these premises. However, my focus will be on rejecting the firstâ&#x20AC;&#x201D;I argue that dreaming does not involve percepts. Before I outline why we ought to reject this premise, it is necessary to understand the motivation behind it. Hobson, a defender of the hallucinatory account of dreaming has supported his position with reference to the neural activity that is reported to occur during dreams and during waking perception. The hallucinatory nature of dreams can be accounted for by the activation of motor and visual areas of the brain during REM sleep. These processes occur in the same way when we have waking perception, illustrating why dreaming (assuming dreaming takes place in REM states)1 involves percepts. More specifically, dreaming involves activity from the brainstemâ&#x20AC;&#x201D;the emission of PGO waves from the brainstem (Sutton 2009: 530). Dreaming is therefore a basic, neurobiological function, just as perception is. The brainstem is responsible for both of these capacities, which suggests why wakeful consciousness and dreams have (arguably) the same 1 Hobson

argues that the brain activity which produces dreaming is a maximal during REM sleep (Pace-Schott and Hobson 2002: 691).

Willa Saadat

105

phenomenal character.2 It is worth noting that when we are awake and having perceptual experience, the stimulation of the relevant cortices (for example the visual cortex) is not the result of the emission of PGO waves from the brainstem, but rather activity that results from stimulation of our sensory organs. Hobson notes that the only difference between the sensory experience one has when awake and the sensory experience one has when they are dreaming is merely the dependence of dreams on internal signal generation. However, there is reason to believe that dreaming is not simply the product of activity from the brainstem and thereby differs from perception. Instead, dreaming is a complex cognitive skill, not a neurobiological given, but instead is something that must be developed over time. A study that supports this conclusion was conducted by Foulkes and his co-workers who inquired into the nature, content and frequency of children’s dreams (Sutton 2009: 533). The study found that children dreamed very little and the dream reports conducted contained descriptions of very brief and insubstantial dreams amongst children aged between three and five years old. Their dreams had very little narrative, emotional range and were infrequent. The suggestion from Foulkes is that this reflects the skill required to be able to dream. Dreaming, as a form of intelligent behaviour, has cognitive prerequisites (Sutton 2009: 534). Hobson’s belief that dreaming is simply a product of brainstem activity does not seem viable under Foulkes’ findings. Instead, this study into children’s dreams suggests that there is a significant difference between dreaming and having percepts. The capacity to have percepts is undeniably a basic neurobiological one whereas the capacity to dream is a skill that must be developed over time. The gradual increase in complexity and frequency of children’s dreams that Foulkes reported suggests that the capacity for dreaming develops in conjunction with the imaginative capacities of individuals. This is further supported by reports that show that the results of tests of children’s imaginative ability correlate with the complexity and frequency of the children’s dreams (Ichikawa 2009: 109). Those who did not perform well in the imaginative ability tests often had simple and infrequent dreams. Imagination must, at the minimum, play an important role in one’s ability to dream—if not constitute one’s dream. This evidence from Foulkes supports a different model of dreams from Hobson’s hallucinatory account, instead it provides reason to accept the view that dreams are imaginative experiences. On this view, dreaming is imagistic and does not involve per2 I say arguably because following from the kind of argument that Hume offers concerning the dis-

tinction between impressions and ideas, some deny the vivacity of dreams in contrast to that of waking perception. See McGinn 2005 for a critique of the Humean distinction between impressions and ideas.

106

Against Dreaming as Hallucination

cepts. Instead of having a (false) perceptual experience when one dreams, the agent is simply imagining the content of their dream. The difference between a percept and an image is that images are ‘not informative’ while percepts are. What McGinn (2005: 18), who offers this characterization, means by this is that when one has a sensory experience, they are receiving information from a cause independent of their own intention—there is something being revealed to them by a stimulus. In other words, McGinn argues that images and imagination is subject to the will and that perception is not. When we imagine something, we do not learn anything from the nature of or the properties of that which we imagine. In contrast, when we perceive something, since we are not the originator of our perceptions, we take in new information that is channelled by percepts. Imagination is a process subject to the will that simulates perceptual experience—it involves agents replicating or constructing what it would be like to see or experience something without actually doing so (Ichikawa 2009: 104–5). The scientific data concerning the correspondence between neural activity and dreams is inconclusive, however, when analysed in conjunction with the above characterization of what images are, it seems to support the view that dreams are imaginative experiences. Zeidman and Maguire have found that cortical activity when we have sense experiences is very similar as that when we are imaginative experiences (Zeidman and Maguire 2016: 179-180). The hippocampus—a brain structure responsible for cognition and memory—plays a role in both visual perception and in imagination. So, based on neurobiology alone, it cannot be determined whether dreaming involves percepts or merely imagination since there is a significant overlap in the regions of the brain activated during each process. But, if we accept the imagination model and Ichikawa’s claim that images simulate percepts, then the observed similarity in cortical activity when imagining and perceiving can be explained by this. The point of visual imagery is simply to enter into a state which has the kind of phenomenal character that perceptual experiences have. (Ichikawa 2009: 106). This explains the similarity in experiences of one’s visual perceptual and one’s visual imagination—the latter mimics the former. Zeidman and Maguire’s conclusions on the function of the hippocampus and its role both in stimulating imagination and perception supports this conceptual tie between these two processes. A crucial question that emerges from the imagination model of dreaming concerns the voluntary nature of images and the involuntary nature of dreams. If the key difference between percepts and images is the latter being subject to the will, then how can dreaming be an imaginative experience when it seems that our dreams are typically outside of our control (Ichikawa 2009: 115–6)? Dreaming tends to involve someone

Willa Saadat

107

having experiences that the agent has not contrived themselves. There is an obvious exception with the case of lucid dreaming, where the agent may be able to exercize control over the content of their dream experience. Ichikawa references the possibility of having lucid dreams to reject the view that dreams contain percepts, due to the lack of control we have over percepts (Ichikawa 2009: 116). However, it seems to be a mistake to make a claim about the nature of dreams in general from evidence derived from cases of lucid dreaming. Lucid dreaming is an exception from the norm and has features that most ordinary dreams do not—it is distinguished by the agent’s recognition that they are dreaming which is clearly not the norm for most dreams. To draw any conclusions on the possibility of voluntary control within all dreams based on the nature of experiences within lucid dreaming would be a hasty overgeneralization. This is not to conclusively suggest that lucid dreams have a significantly different character from non-lucid dreams, rather we simply cannot assume that they are. More needs to be said on the similarity between lucid and non-lucid dreams before we can generalize from one to the other. So, with the exception of lucid dreaming, if we are committed to the view that dreams are involuntary and imagination is voluntary, then it seems natural to argue that dreaming involves percepts which are also involuntary. The defence of the imaginative model relies on a distinction between voluntary control and subject of the will. Imaginative experiences can be a product of the will and not under our voluntary control (in waking consciousness, imaginative experiences seem to be both however exercising control is not necessary for one to have an imaginative experience). As Ichikawa puts it, not being able to control whether or not I do something, or imagine something, does not mean that it was not me who did or imagined that (Ichikawa 2009: 116). We can accept that dreams are involuntary, but we can also maintain that they are still subject to the will of the agent—it is still a product of the agent’s doing. Likewise, imagination is subject to the will and thereby a product of the agent’s doing. The narrative that dreams often possess indicates that this is the case. Seeing as most dreams have a degree of coherence, in that there seems to be a connection between events or a beginning, middle and end—we have reason to suppose that this narrative emerges from being subject to the will of the dreamer (McGinn 2005: 84). Dreams have meaning and can convey a sequence because we have designed them. There is some kind of agency or will being displayed here without signs of explicit voluntary control. If dreams were entirely independent of the will, then it becomes difficult to account for how they would come to have a coherent structure or narrative. Most dreams contain more than just the experience of various random

108

Against Dreaming as Hallucination

images—the narrative that dreams have must originate from somewhere and percepts are not capable of fulfilling this function. Instead, the dreamer creates this narrative that their dream has since they are simply imagining the content of their dream. From the points outlined in this essay, it seems that there is a strong reason to accept the view that dreams are imaginative experiences and deny the hallucination model. This rejection of the hallucinatory nature of dreams rests on the denial of the involvement of percepts when we are dreaming. Empirical studies into children’s dreams and the connection between imaginative capacities and the frequencies and complexity of dreams suggests that dreaming requires a more developed skill, namely imagination, as opposed to perception which is a basic neurobiological capacity. Although there are prima facie concerns about the inconsistency between the involuntary nature of dreams and the voluntary nature of imaginative experiences, these can be overcome with a more qualified account of what imaginative experiences are and how they do not require one to exercize direct control, rather they are simply subject to an individual’s will. From these considerations, dreaming does not appear to involve percepts and we therefore have reason to deny that dreams are hallucinations. King’s College, London, England

References Hobson, J. A. (1999). Dreaming as delirium: How the brain goes out of its mind. MIT Press. Ichikawa, J. (2009). Dreaming and imagination. Mind and Language 24(1), 103–121. McGinn, C. (2005). Mindsight: Image, Dream, Meaning. Harvard. Pace-Schott, E. F. and J. A. Hobson (2002). The neurobiology of sleep: Genetics, cellular physiology and subcortical networks. Nature reviews neuroscience 3(8), 591– 605. Sutton, J. (2009). Dreaming. In J. Symons and P. Calvo (Eds.), The Routledge Companion to Philosophy of Psychology. Routledge. Zeidman, P. and E. A. Maguire (2016). Anterior hippocampus: The anatomy of perception, imagination and episodic memory. Nature Reviews Neuroscience 17(3), 173– 182.

Recognition in Hegel’s Phenomenology jared smith

1 Introduction NE of the most renowned and influential claims of Hegel’s Phenomenology of Spirit is that ‘[s]elf-consciousness exists [. . . ] only in being recognized [als ein Anerkanntes]’ (PdG §178, emended). Hegel, however, by and large introduces the concept of ‘recognition [Anerkennen]’ without giving a definition; and it is thereby not immediately clear what exactly recognition is, nor why it is necessary for self-consciousness. Some commentators suggest that, for Hegel, an act of recognition is, tout court, an act of ‘self-negation’ (PdG §175) whereby the recognizing self-consciousness disregards itself so that it is ‘for the other [sc. the recognized self-consciousness]’ (PdG §175).1 In my view, while this reading of ‘self-negation’ explains why the self-negation of one self-consciousness mediates the self-certainty of another, the suggestion that recognition just is self-negation requires qualification. Selfconsciousness, says Hegel, acquires an explicit certainty of the ‘equality [Gleichheit] of itself with itself’ (PdG §167, emended) by being recognized. Yet the self-negation of one self-consciousness only allows another to acquire an ‘ambiguous [doppelsinnige]’ (? ) self-certainty by which it is certain of itself as something other than it, and is thereby not recognition. On account of this, I argue, the movement by which the recognized self-consciousness ‘supersede[s] [aufheben] this otherness of itself’ (PdG §180) to acquire a certainty of itself as ‘equal [gleich] to itself’ (PdG §181) is equally constitutive of recognition. My discussion proceeds as follows. First, I outline Hegel’s account of self-consciousness as ‘Desire [Begierde]’ (PdG §167) (§2). I further explain why the experience of Desire results in the duplication of self-consciousness (§3). Subsequently, I demonstrate why the self-negation of one self-consciousness mediates the self-certainty of another (§4).

1 See,

e.g., Houlgate 2013: 85–93, especially 88–89. This reading of ‘self-negation’ is Hans-Georg Gadamer’s (1976: 61).

British Undergraduate Philosophy Review 1 (2020)

109

110

Hegel on Recognition

I then argue that, for Hegel, recognition consists equally in two movements (§5). Finally, I explain why, in Hegel’s view, recognition is essentially the double-sided action of two self-consciousnesses (§6).

2 Self-consciousness as ‘desire [ Begierde]’ Much of what Hegel says in chapter four of the Phenomenology (‘Self-consciousness’ [§166-230]) presupposes that self-consciousness has a ‘double object’ (PdG §166): first of all, itself; and secondly, something other than it (e.g., an object of sense-certainty or perception, or another self-consciousness). While almost anyone would agree that self-consciousness has itself as its object (it is self-consciousness, after all), it is less clear why self-consciousness also has as its object something other than it. This point, then, requires some clarification. In the preceding chapter (‘Consciousness’ [§190-165]), Hegel explains how the experience of consciousness gives rise to self-consciousness as a new shape of consciousness when ‘[w]hat the object [of consciousness] immediately was in itself’ turns out to be nothing more than ‘a mode in which the object is only for another’.2 (PdG §166) On the one hand, Hegel claims, this other that was the object of consciousness has ‘vanished’ (PdG §167); on the other, we are told, the respective moments of this other are preserved for self-consciousness. On this account, self-consciousness does not only have itself as its object: it confronts a ‘sensuous world’ (PdG §167) of perceivable objects which exists only as ‘appearance [Erscheinung]’ (PdG §165)—viz., as something which is nothing in itself and is only there for self-consciousness. Self-consciousness, however, must become explicitly certain of itself: the ‘unity [of self-consciousness with itself] must become essential to self-consciousness’ (PdG §167). In other words, self-consciousness must ‘return from otherness [Rückkehr aus dem Anderssein]’ (PdG §167) so that it has only itself as its object. It does so, Hegel claims, by removing the antithesis between the sensuous world or appearance, which it confronts qua consciousness, and the ‘unity of self-consciousness with itself’ which is the ‘truth’ for self-consciousness (PdG §167). Hegel calls this movement by which the antithesis between appearance and truth is removed ‘Desire [Begierde]’ (PdG §167). Moreover—since self-consciousness is, namely, that by which explicit self-certainty is acquired—it follows that self-consciousness just is this movement of Desire: ‘[S]elfconsciousness is Desire in general [Begierde überhaupt]’ (PdG §167). Self-consciousness, 2 Of course, this is a vast oversimplification of the previous chapter of the Phenomenology. See Houlgate 2013: 31-82 for an extensive discussion.

Jared Smith

111

as Desire, seizes upon and destroys the independence of the other which it confronts; by doing so, it removes the antithesis between itself and this other and thereby ‘gives itself the certainty of itself as a true certainty’ (PdG §174). The action of Desire might be crudely illustrated as follows: Desire confronts an object (e.g., an apple), destroys this object by consuming it, and, in doing so, restores explicit self-certainty.3 Self-consciousness is therefore not a state in which the ‘I’ is simultaneously the content of ‘the relation and the relating itself [der Beziehung und das Beziehen selbst]’ (PdG §166, emended). For a state of this configuration, where the content of the ‘relation’ and the ‘relating’ is the same, there can be no movement (something cannot, from itself, return to itself ). Such a configuration of self-consciousness would therefore merely be the ‘motionless tautology of: I am I’ (PdG §167). Instead, self-consciousness is a movement: it is a return from sensuous otherness.

3 Another self-consciousness I next explain why Hegel claims that the experience of Desire impels self-consciousness to confront another self-consciousness. As Desire, we are told, the object for self-consciousness ‘has the character of a negative’ (PdG §167): the object is regarded as having no independent existence of its own and thereby as something that merely exists for self-consciousness—viz., as something that is merely there to be consumed in order to restore explicit self-certainty. As such, self-consciousness takes itself to be the only thing that is essential and, conversely, the object to be wholly inessential. Hegel claims, however, that self-consciousness learns in its experience that it must confront a substantive, independent other. As Hegel says, ‘Desire and the self-certainty obtained in its gratification, are conditioned by the object, for self-certainty comes from superseding this other: in order that this supersession can take place, there must be this other’ (PdG §175, emphasis own). The very attempt of Desire to destroy its object, in other words, just affirms that there is this other, which is essential for acquiring self-certainty, and which self-consciousness is thereby ‘unable to supersede’ (PdG §175). Desire, of course, may successfully destroy particular objects by consuming them (e.g., eating apples and pears); however, it can never eliminate otherness entirely. Accordingly, self-consciousness learns that it is dependent on the object, which is, in turn, independent of it—notably, the inverse what 3 As Robert R. Williams (1992: 145) points out, Desire is metaphysical: ‘not only does it show the nul-

lity of finite objects (e.g., fruit), it brings them to immediate presence, and then devours and annihilates presence’.

Hegel on Recognition

112

self-consciousness immediately took the relation between itself and its object to be. Yet, Hegel claims, self-consciousness ‘must experience satisfaction, for it is the truth’ (PdG §174, emphasis own), and it can only do so by superseding the object. How, then, does self-consciousness experience satisfaction and thereby acquire explicit self-certainty when it cannot supersede the object? It does so, we are told, when the object ‘effects the negation within itself’ (PdG §174) and still remains independent— viz., when the object of consciousness is one which, ‘of its own self, posits its otherness or difference as a nothingness, and in doing so is independent’ (PdG §176). Now, in Hegel’s view, the only object that can beget its own negation in a way by which it retains its independence is another self-consciousness. Only consciousness can renounce each and every feature by which it is characterized and still remain aware of itself as a pure ‘I’—or, as Hegel says, remain ‘for itself a genus [er für sich selbst Gattung]’ (PdG §176). In so far as consciousness reduces itself to nothing and remains abstractly self -conscious, it maintains an awareness of its own independent existence; it remains, in other words, ‘equally independent in this negativity of itself’ (PdG §176). Hegel thereby concludes, if self-consciousness is to acquire explicit self-certainty, it must confront another self-consciousness. In this way, the experience of Desire impels self-consciousness to confront another self-consciousness.

4 Self-negation and self-certainty The experience of Desire, then, results in the ‘duplication of self-consciousness’ (PdG §176). However, it is not yet entirely clear why the fact that one self-consciousness confronts another that is ‘for itself a genus’ enables the former to acquire explicit selfcertainty: The mere fact that there is another self-consciousness—existing out there, so to speak—does not explain how the first self-consciousness comes to relate only to itself. As Houlgate remarks, ‘[that there is another self-consciousness which is “for itself a genus”] allows the first self to see in the other is another self-consciousness. It does not allow it to see only itself in the other’ (Houlgate 2013: 88). As such, an alternative reading of ‘self-negation’ is needed that explains why the self-negation of one self-consciousness mediates the self-certainty of another. Such a reading of ‘self-negation’ is provided by Hans-Georg Gadamer, who claims the following: ‘[Self-consciousness] is able to find itself in the other if this other is independent and grants that it does not exist in its own right, but rather that, in disregard of itself, it “is for another” [“für das Andere zu sein”]’ (Gadamer 1976: 61).4 4 Note

that the ensuing reading is only briefly suggested by Gadamer; it is made more explicit by

Jared Smith

113

On this reading, self-negation is not merely an act whereby one self-consciousness renounces each and every feature in order to conceive of itself as a pure ‘I’; it is an act whereby one self-consciousness disregards itself in such a way that it exists completely ‘for another’ (PdG §175, emended). In so far as one self-consciousness is ‘for another’ self-consciousness, the latter sees only itself in the former—viz., it ‘does not see the other [sc. the former self-consciousness] as an essential being, but in the other sees its own self ’ (PdG §179, emphasis own). Only by interpreting ‘self-negation’ in this way, in my view, can we grasp why the self-negation of one self-consciousness mediates the self-certainty of another.

5 ‘Recognition [ Anerkennen]’ Now, Hegel begins section A of chapter four of the Phenomenology by claiming that ‘[s]elf-consciousness exists in and for itself [. . . ] only in being recognized [als ein Anerkanntes]’ (PdG §178, emended). By this, Hegel means that self-consciousness is only explicitly certain of itself when it is recognized by another. Hegel, however, largely introduces the concept of ‘recognition [Anerkennen]’ (PdG §178) without giving a definition; and consequently, it is not immediately clear what exactly recognition is, nor why it is necessary for acquiring explicit self-certainty. I now argue that, for Hegel, recognition consists equally in two movements. First, I would like to mention another interpretation of recognition. Some commentators suggest that, for Hegel, an act of recognition is, tout court, an act of selfnegation whereby the recognizing self-consciousness disregards itself so that it is ‘for the other [sc. the recognised self-consciousness]’ (PdG §175). For example, Houlgate (2013: 89) suggests that an act of self-negation, whereby one self-consciousness turns itself into a ‘mirror’ for another, just is an act by which the former recognizes the latter.5 Houlgate (2013: 89) writes, ‘I recognize another, therefore, simply by turning myself into a mirror for that other’. Yet, in my view, while self-negation is certainly constitutive of recognition, it is only part of the story. One self-consciousness, indeed, acquires self-certainty in so far as it ‘finds itself’ (PdG §179) in a second self-consciousness which, in disregard of itself, is for the first. However, Hegel maintains, the first self-consciousness has at the same time ‘lost itself’ because it finds itself as an ‘other being’ (PdG §179). The outcome, we are told, is a self-certainty that is ‘ambiguous [doppelsinnige]’ (PdG §181): on the one Houlgate (2013: 88–89). 5 See also, e.g., Pinkard 1994: 52.

Hegel on Recognition

114

hand, the first self-consciousness acquires a certainty of itself; on the other, however, it is certain of itself as something other than it — viz., it is certain of itself as something that exists out there. — Hence it is not certain of the ‘equality [Gleichheit] of itself with itself’ (PdG §167, emended). By itself, the self-negation of one self-consciousness is thereby deficient for restoring explicit self-certainty, and is not recognition. For the first self-consciousness to acquire an explicit self-certainty, another movement is necessary: the first self-consciousness must ‘supersede [aufheben] this otherness [Anderssein] of itself’ (PdG §180). Only thereby, we are told, can the first selfconsciousness acquire an explicit self-certainty by which it is certain of itself as ‘equal [gleich] to itself’ (PdG §181). However, Hegel says, this supersession is itself ambiguous: for, by superseding the second self-consciousness to become ‘certain of itself as the essential being’ (PdG §180), the first self-consciousness, we are told, ‘proceeds to supersede its own self’ (PdG §180). The first acquires a certainty of itself in the second; by superseding the second, it loses this certainty. Hegel claims, however, that by superseding its otherness the first self-consciousness now acquires an explicit self-certainty. He writes, ‘[t]his ambiguous supersession of its ambiguous otherness is equally an ambiguous return into itself [Ruckkehr in sich]’ (PdG §181). On the one hand, by superseding its otherness, the first self-consciousness ‘receives back its own self’ (PdG §181) and in so doing acquires an explicit certainty of itself as ‘equal [gleich] to itself’. On the other hand, the ﬁrst self-consciousness ‘gives the other self-consciousness back to itself [. . . ] and lets the other again go free’ (PdG §181, emended).6 The second self-consciousness is thus not destroyed by the first’s supersession; it is instead preserved and granted the freedom to give or withhold recognition of its own accord.7 So, for Hegel, recognition does not only consist in the movement of self-negation: The movement whereby the recognized self-consciousness supersedes its otherness in order to acquire a certainty of itself as ‘equal [gleich] to itself’ is equally constitutive of recognition. 6 See Williams 1992: 155 for a discussion of the translation of this passage. 7 Houlgate (2013:

91) and Gadamer (1976: 61), among others, suggest that, in letting the second selfconsciousness go free, the first recognizes the second. In my view, this misrepresents Hegel’s argument. In being let go free by the first, the second is ‘allowed to be’ a genuinely independent self-consciousness; however, this does not mean that the first recognizes the second as one.

Jared Smith

115

6 Mutual recognition Finally, I briefly explain why Hegel claims that recognition is essentially the doublesided ‘action [Tun]’ (PdG §182) of two self-consciousnesses. One way by which recognition differs from Desire is that the object of self-consciousness no longer exists simply for self-consciousness: the object ‘has an independent existence of its own’ (PdG §182)—viz., it is another self-consciousness. As such, in the case of recognition, selfconsciousness has as its object one which also ‘exists only in being recognized’ (PdG §178)—viz., the object of self-consciousness is one which is only ‘for itself a genus’ (PdG §176) in so far as it is recognized by another self-consciousness. The second selfconsciousness, for it to recognize the first, must thereby be recognized by the first, and vice versa. Hegel accordingly writes, ‘the first may not use the other for its own ends, unless the other does for itself what the first does’ (Williams 1992: 156 (§182)). Each selfconsciousness must disregard itself so that it is for the other, and each must supersede its otherness to acquire an explicit self-certainty. In this way, recognition necessarily involves the joint action of both self-consciousnesses: ‘[I]t is indivisibly the action of one as well as of the other’ (PdG §183).

7 Conclusion In conclusion, in this essay I have shown why, for Hegel, acquiring explicit self-certainty requires mutual recognition. I first explained why self-consciousness immediately takes the form of Desire and why, in the experience of Desire, self-consciousness confronts another self-consciousness. I then outlined why the self-negation of one self-consciousness mediates the self-certainty of another. I further argued that, for Hegel, recognition consists equally in two movements: firstly, the recognizing self-consciousness disregards itself so that it is ‘for the other [sc. the recognized self-consciousness]’ (PdG §175); secondly, the recognized self-consciousness supersedes its otherness in order to acquire a certainty of itself as ‘equal [gleich] to itself’ (PdG §181). Finally, I explained why recognition essentially takes the form of mutual recognition. University of Warwick, Coventry, England

References Gadamer, H. (1976). Hegel’s Dialectic: Five Hermeneutical Studies. Yale.

116

Hegel on Recognition

Hegel, G. W. F. (1977). Phenomenology of Spirit. Oxford. [Ger. Phänomenolgie des Geistes, trans. A. V. Miller]. Houlgate, S. (2013). Hegel’s ‘Phenomenology of Spirit’: A Reader’s Guide. Bloomsbury. Pinkard, T. (1994). Hegel’s Phenomenology: The Sociality of Reason. Cambridge. Williams, R. R. (1992). Recognition: Fichte and Hegel on the Other. SUNY Press.

pgfexternal@did@a@shipout