Human Digital Identity and AI

id191831   HUMAN DIGITAL IDENTITY AND AI BERTA FITÓ CASAS Thesis supervisor ULISESCORTÉSGARCÍA(BSC) Thesis co-supervisor ATIACORTÉSMARTÍNEZ(BarcelonaSupercomputingCenter) Degree Master'sDegreeinArtificialIntelligence Master's thesis School of Engineering Universitat Rovira i Virgili (URV) Faculty of Mathematics Universitat de Barcelona (UB) Barcelona School of Informatics (FIB) Universitat Politècnica de Catalunya (UPC) - BarcelonaTech 27/01/2025  Acknowledgments I would like to express my deepest gratitude to all the people and organisations who have accompanied me throughout this journey. First, my supervisors, Ulises Cort´es, for his insightful and challenging views, and Atia Cort´es Mart´ınez, for her valuable feedback —and to both of them for their invaluable guidance and support in completing the significant undertaking of this thesis research. I would also like to thank the Universitat Polit`ecnica de Catalunya, Universitat de Barcelona, and Universitat Rovira i Virgili for making it possible for me to enrol in this Master’s Programme. Additionally, I extend my gratitude to the Barcelona Supercomputing Center – Centro Nacional de Supercomputaci´on (BSC–CNS) for providing the resources and facilities that enabled me to complete and sustain progress on this project; specifically, to the CS@BSC department for the grant of curricular credits. Finally, I would like to convey my heartfelt gratitude (and slight apology) to my family, friends, and everyone who endured my endless rants about my latest readings on identity to the point that we all became attuned to finding it everywhere. Special thanks to those who provided books and articles they thought might contribute to this thesis —and to the few who persevered in reading it through to the end. i ii Contents Contents Acknowledgments i Abstract v 1 Introduction 1 1.1 Definition and etymology of identity ......................... 2 1.2 What is an Autonomous Agent? . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.3 Motivation ....................................... 10 1.4 Hypotheses and Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 1.5 Documentstructure .................................. 13 2 Analytical Identity 17 2.1 Leibniz’sLaw...................................... 20 2.2 Relativeidentity .................................... 25 2.3 Compositionasidentity ................................ 33 2.4 Contingentidentity................................... 39 2.5 Identityacrossworlds ................................. 41 2.6 Identityacrosstime .................................. 44 2.7 Vagueidentity ..................................... 48 iii iv Contents 2.8 TheFleetofTheseus.................................. 50 3 Personal Identity 53 3.1 Brain duplication against personal identity . . . . . . . . . . . . . . . . . . . . . 64 3.2 Psychologists are also getting involved . . . . . . . . . . . . . . . . . . . . . . . . 72 3.3 What memory says about us . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 3.4 Our place among the animals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 4 Cultural Identity 103 4.1 Social and political identity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 4.2 Thestoryofourlives.................................. 117 4.3 Artimitateslife..................................... 121 5 Legal Identity 131 5.1 Legislative triangulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 5.2 The net of identity legislation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 5.3 Biometrics, anonymity, and robot rights . . . . . . . . . . . . . . . . . . . . . . . 148 5.4 Digitalescalation.................................... 155 6 Agent Identity 163 6.1 What is the digital world? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165 6.2 The double-edged sword of online anonymity . . . . . . . . . . . . . . . . . . . . 168 6.3 Definition of digital identity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171 7 Conclusions 179 7.1 Futurework....................................... 186 Abstract This thesis aims to establish whether the traditional view of identity —specifically, personal identity— can be applied to Autonomous Agents based on the BDI (Belief-Desire-Intention) model. Its purpose is to provide a theoretical platform that may serve as a foundation for further discussions on the role intelligent machines might occupy in the future, as the digital world becomes increasingly embedded into our daily lives. With Artificial Intelligence systems integrating in human social and legal frameworks, defining their identity holds significant ethical and practical implications. To this end, the thesis presents a comprehensive overview of the historical discourse on identity across four interwoven areas of progressive abstraction: analytical philosophy, philosophical discourse on personal identity, sociocultural studies, and legislation. It proposes a novel framework that synthesises insights from these disciplines. Each field provides tools for forming a fuller and more informed conclusion on the subject, which are ultimately applied to a proposed definition of identity capable of accommodating both human and non-human entities. A key concept in this discussion is the distinction between self and identity, terms that have often been conflated or confused. v vi Contents Chapter 1 Introduction Our identity is something we deal with and trade in on a daily basis, from using plastic cards to borrow books or hire movies to numerical representations of that identity in PINs to the production of formal certification of our identity. (Jon Austin in Culture & Identity, Identity & Identity Formation) Identity has never known consensus. Western and non-Western doctrines take seemingly irreconcilable views: one focuses on the one, the other on the many. Each side, in turn, is deeply conflicted from within. In Western discourse, some reject there even being a need for debating on the matter at all, with voices such as Peter Thomas Geach’s, who led the movement denying the existence of an absolute account of identity (i.e., a match in all possible characteristics of an object) in favour of relative identity relations. David Kellogg Lewis, one of the most important analytical philosophers of the 20th century, thought of it as an utterly unproblematic notion, arguing that no philosophical problems concern identity (Noonan and Curtis [2022]). Willard Van Orman Quine saw identity as such a basic concept that we do not have a definition for it, so it becomes tough to tackle challenges that may arise from its study. Defining identity is a global challenge. Western thinking, deeply influenced by JudeoChristian ideologies, focuses more on the individual, which has narrowed its search for a definition of identity to a particular object, contrasting Eastern thinkers’ more communal or societal focus. This leads to topical issues such as the harmonisation of Artificial Intelligence legislation, 1 2Chapter 1. Introduction the boundaries of which should be influenced by the different definitions of a person’s identity according to each culture and prevailing philosophy. To attempt to minimise a more than probable Western bias and foster an as universal as possible view on the matter at hand, this work will examine the definition of identity through four layers (analytical identity, philosophy on personal identity, cultural identity, and legal identity) before eventually reaching a proposal on the understanding and definition of agent identity —the ultimate focus of this thesis. Although the analysis of the first two layers will heavily feature Western ideas, they should be contrasted with as many non-Western proposals on the same topic as possible. 1.1 Definition and etymology of identity When thinking of identity, one’s mind will most often go to the concept of personal identity. Outside philosophy, according to Olson [2023], it encapsulates the qualities to which we feel the most attached, which we consider to define us as a person or make us the person we are. From this follows what might be construed as cultural identity, in which a person achieves full humanity in an accepted context of symbols, values and behaviours within a community. This is one of the most significant contemporary challenges to many Western philosophical claims about the person and society (Loughney [1998]). However, the notion of identity is also present in analytic philosophy and mathematics. In logic and metaphysics, identity and sameness mean the same —or, in other words, their meanings are identical. However, there are nuances to consider. A distinction is usually made between qualitative and numerical identity. Things with qualitative identity share properties so that they can be more or less qualitatively identical. In contrast, numerical identity is a relation between a thing and no other (Noonan and Curtis [2022]). Throughout this thesis —and especially in §2— we will focus on numerical identity, which is usually defined as the equivalence relation that satisfies Leibniz’s Law (see §2.1), as long as what is true of xis understood to include being identical to x—otherwise, Noonan and Curtis [2022] assert, the claim would be too weak. 1.2. What is an Autonomous Agent? 9 counterfeit people: artificial avatars which pass for real people in digital environments. These counterfeit people are the most dangerous artifacts in human history, capable of destroying not just economies but human freedom itself. (Daniel C. Dennett in The Problem With Counterfeit People) Like Go˜ni, Dennett [2023] sounds the alarm about our natural tendency to treat as a person anything that seems to fit our understanding of a human being —citing the ability to “talk sensibly”. He fears that these replicas’ capability to confuse us and exploit our fears and anxieties will lead us into temptation and, from there, into yielding to our subjugation. He supports philosopher and historian Yuval Noah Harari’s call for Artificial Intelligence to be obligated to disclose that it is Artificial Intelligence. Though not yet stained by first-hand dread of the future, this kind of discussion over the capabilities and limitations of Artificial Intelligence is not new. Searle [1980] argued that computers could at best simulate, but never possess intelligence —machines could never truly understand natural language nor have propositional attitudes, such as beliefs, even if they were eventually able to pass the Turing Test. This would seem to complicate the endeavour of this thesis, but it is an argument that relies heavily on the intrinsic nature of what fixes an identity, which the reader will ultimately find contested in these pages. Searle [1980] presented an argument known as the Chinese Room, which posits that a computer program that follows instructions to respond to messages does not really understand the language in which they are written, just as Searle himself would not understand Chinese by following a manual to match input characters with appropriate responses. In response, Cole [1991] developed the notion of virtual person. He thought Searle [1980]’s argument was logically invalid because it lacked a form of personal identity theory inspired by the likes of John Locke, Derek Parfit, John Perry and David Lewis, all of whom will feature prominently in this thesis. Cole [1991] admits that Searle [1980] successfully proved that no computer will ever understand any natural language but gets entangled in defending a sort of emergent self within both minds and machines: [M]inds and persons are not the same as the machines, biological or electronic, that 10 Chapter 1. Introduction realize them. (David Cole in Artificial Intelligence and Personal Identity, Abstract) He argues that the machine itself not being able to understand natural language does not show that nothing does, and thus that Searle [1980] failed to establish any limitations on Artificial Intelligence. According to Cole [1991], Searle’s activity in receiving the notes in Chinese, following the instructions to answer them, and sending back those answers, even though he could neither understand nor express anything about himself on them, could realise a virtual mind, as if a machine were emulating another one. As the thesis unfolds, his proposal will be enriched and contested by many others. 1.3 Motivation Humans have always been concerned with defining our own identity. First, we understood us as different from them —other forms of living beings, animals or other groups of people. Then we began individualising, looking for ways to identify me against you. And yet, the isolation of human portraits will never be found in a cave painting —it is much newer. The same can be seen in literature, with the peppered use of the first person as a narrative tool in legendary works like the Epic of Gilgamesh (c. 2000 BCE3) and Homer’s Odyssey (8th century BCE), although it did not become commonplace until the 16th century with novels like Daniel Defoe’s Robinson Crusoe (1719) and Jonathan Swift’s Gulliver’s Travels (1726). The increasing withdrawal from the many is evident in the development of art —which may be due to a dwindling identification with the surrounding nature and our tribe. We find the first recorded abstract reflection on identity more than two thousand years ago, during Athens’ democracy, commonly known as its Golden Age (5th-4th centuries BCE). Barnes [1977] observes that some scholars believe that Leibniz’s Law, a concept that will be discussed extensively as we delve into this work, is due to Aristotle’s remarks on numerical equality in Topica, written during that time. Although it is impossible to know what earlier thought on 3Before Common Era. It is a secular term used to denote years before the start of the Common Era (CE), which corresponds to the years traditionally counted from the birth of Jesus Christ. 1.3. Motivation 11 identity may have been lost to time, we can follow the blossoming of the analytical, philosophical, cultural, and legal debate on identity that has sprung from these first observations to the present day. Some (but by no means all) notable Western names include Parmenides, Heraclitus, Plato, John Locke, Ren´e Descartes, Georg Wilhelm Friedrich Hegel, David Hume, Hannah Arendt... who were either remarkably close in some respects (such as Hume to Vasubandhu, as Hansen [2016] points out) or differed significantly from non-Western thinkers. Siddhartha Gautama (also known as the Buddha, the Enlightened One) stands out among the latter —he taught a doctrine involving the self about two centuries before Aristotle and was roughly coetaneous with Heraclitus. A notable chasm separates Western from non-Western thought: the former clasps a kind of individualism that aspires to define the person by itself. At the same time, the latter sees the individual as an extension of the intricate web of relationships within a community. As we will more prominently see in §4, this fundamental difference has led to two understandings of identity that have so far very rarely intersected. As listed by Noonan and Curtis [2022], recent work on the matter of identity has been concerned with the notion of a criterion of identity4and the proper analysis of identity over time (in particular, the disagreement between advocates of perdurance and endurance), among other topics which we will briefly introduce throughout these pages. Other areas of study concern the definition of personal identity, which has been so far dominated by Western debate over the existence of a key component to personhood —whether it may be the brain, the mind, memory, consciousness, or the soul. The effect of community and culture is being analysed in an area that has begun to see more convergence between Western and non-Western voices in recent decades. One word can summarise the state of the field of identity: fragmented. There is little to no accord between theories and proposed solutions on the polemical issue of its definition —which is to be expected, given that it is something that is very much a matter of concern to us all. This work does not aspire to mend it, as it has been a kaleidoscope in the making for millennia. 4The term criterion of identity is ambiguous. In an epistemological sense, it is a way of saying whether an identity proposition is true or a way of recognising whether an individual Ais identical to an individual B. However, it also has a metaphysical interpretation, according to which it is a set of (non-trivial) necessary and sufficient conditions for the truth of an identity proposition (Mackie and Jago [2022]). 12 Chapter 1. Introduction Instead, the motivation for writing the following pages comes from the irruption of a new entity that may need its own meaning for identity. It is no longer the case that humans are the only ones in need of a characterisation —Autonomous Agents, which the field of Artificial Intelligence has designed to emulate us, raise the question of whether the human definition of identity should be expanded to include them if a new one ought to be issued... or if they need not be included in the discussion at all. This work will attempt to link together the ideas on human identity discussed so far to create a coherent, generally acceptable definition of identity that is capable of satisfactorily providing a theoretical framework for the analysis of digital identity, which (broadly stated, see §6 for a more in-depth discussion) encompasses a digital user and all their online facets. Said user may either be a human being or an Autonomous Agent. The project aims to determine whether these virtual identities could be added to this philosophical debate and what this would look like. 1.4 Hypotheses and Objectives The following thesis is built around one central question, which branches out into three others: Q0 What is the definition of identity? Q1 Is identity an intrinsic quality of its object? Q2 Is identity a phenomenon exclusive to humans? Q3 Is identity a relevant issue in the field of Artificial Intelligence? The definition of identity (Q0) becomes the objective of this thesis: our motivation, as stated in §1.3, is to provide a comprehensive explanation of identity that might include that of Autonomous Agents as well as that of humans. The other three questions (Q1, Q2, Q3) we will tentatively answer here, in the shape of hypotheses: H1 Identity is not an intrinsic quality of its object. Rather, it is a construction of all the 1.5. Document structure 13 expressions of a person’s life that may be interpreted both by the individual and by all those around them. H2 Identity is not exclusively human. At the very least, it may be assigned externally, possibly without the subject undertaking the necessary introspection to interpret their own life experience. It is, therefore, more akin to a social product. H3 As the world becomes increasingly digital, the need for a clear and comprehensive definition of identity becomes more relevant —especially one that encompasses the digital realm. This leads to a profound discussion about a person’s digital profiles and footprint, as well as their interaction with an entity that may also leave a digital trail: Autonomous Agents. Concerns about ethics, responsibility for an agent’s actions, and, with the rise of Artificial Intelligence art, authorship are among the current issues that are severely undermined by the lack of a clear consensus on the identity of these entities. Acknowledging this, it follows that identity is a relevant issue in this field. 1.5 Document structure The following chapters will serve as an exhaustive overview of the historical and current discussion on the topic from four different perspectives, all erecting scaffolding to sustain the construction of a fifth view, the objective of this thesis: 1. Analytical identity: This chapter focuses on the mathematical side of identity. It discusses the notion in a very comprehensive manner, questioning the sameness of whichever two objects from a logical standpoint. Debates over the absolutism of identity, as well as its answer to changing times and worlds, are central to it. It is the basis of all subsequent discussions, and the starting point for answering Q1. 2. Personal identity: From this point on, identity is restricted to that of a thinking being, most commonly deemed a person. It is usual in this facet of the discussion to consider identity an intrinsic notion attached to an arbitrary object (e.g., the brain, the soul) whose function is to maintain a continuity in a person’s identity. As a result, psychology and 14 Chapter 1. Introduction philosophy are closely intertwined, as is the discussion of what it means to be human. Thus, the debate adds nuance to Q1, while undertaking Q2. It is in this chapter that a notable issue emerges: the problem of identity5is treated indistinctly for the problem of the self6, thus muddying all debates on the matter. 3. Cultural identity: In this chapter, identity becomes a broader concept, taking into account not only the social influences on identity but also the cultural norms to which it adheres. Q1 and Q2 gain depth. Finally, the individualism of Western thought parts and allows non-Western voices to be heard, critiquing and expanding the concept of identity not only as an individual phenomenon but as a community-wide one. One of the arbitrary objects to which identity (or the self) has been attached in personal identity is particularly important in the discussion of cultural identity: individual and collective memory. To illustrate the pervasive presence of identity in our everyday cultural lives, a sample of some works of pop culture is examined. 4. Legal identity: Once the theoretical background to the definition of identity has been established, its current practical application is examined. In doing so, Q3 begins to take shape. Legislation provides a clear definition of identity in a pragmatic sense: so that its rules on citizenship and authorship, or its criminal law, can be applied. This chapter uses the different approaches of the European Union, the United States of America, and the People’s Republic of China to illustrate the lack of a hegemonic definition, even in practice. It also introduces discussions on Artificial Intelligence regulation, democracy, and identification. 5. Agent identity: Finally, Q0 is answered and a definition of identity is presented which includes both human beings (in the real and digital worlds) and Autonomous Agents. In its elaboration, its previously introduced influences are delimited, so that the path towards the objective of this thesis is clear and provides an easily traceable genealogy of ideas. 6. Conclusions: A survey of the research of this thesis is conducted, and its conclusions are 5The problem of identity concerns that which makes an object the same as another. In personal identity, it is broadly the study of the sameness of one person to another in a different time and space. 6The problem of the self concerns that which makes an object an intrinsically unique and continuous being. It may be limited to personal identity, although there are discussions, dating back at least to the Presocratics, about the nature, or essence, of objects. 1.5. Document structure 15 reviewed in order to verify that its objective has been achieved, as well as having seen its hypotheses either proved or disproved. The short timeframe of this project and the complexity of the issue have made it impossible to include a complete account of the discussion on identity in this document. For this reason, each perspective serves as a sample of a broader topic —yet even if some voices are excluded, the overall contested state of the field transpires as intended and paints a picture that allows the conclusions of this thesis to be reached. Chapter 2 Analytical Identity Identity only becomes an issue when it is in crisis, when something assumed to be fixed, coherent and stable is displaced by the experience of doubt and uncertainty. (Kobena Mercer in Welcome to the Jungle: Identity and Diversity in Postmodern Politics) A comprehension list of proposed definitions of identity and general similarity will be contrasted throughout this chapter. According to logical tradition, Deutsch [1997] writes, both identity and general similarity are reducible to relations of strict identity —i.e., identity as treated in standard logic. On the other hand, personal identity is a branch of analytical metaphysics (Cole [1991]). Relations of identity can be expressed as x and y are the same sortal concept, while relations of general similarity may be read as x and y are (or have) the same property (Deutsch [1997]). Conventionally, identity has been regarded as a purely logical notion, while general similarity has not. This chapter provides a brief introduction to each proposal and some of the controversies in which they are involved. Although not the focus of this thesis, as its overarching aim is identity as applied to a subject (human or otherwise), these are the substrates or tools with which theories of personal identity are built. The classical (and most conservative) definition for identity is known as absolute identity. It makes it so for one object to be identical to another, all its properties and relations must be the same (Noonan and Curtis [2022]). This idea often leads to the conclusion that the only thing an 17 18 Chapter 2. Analytical Identity object is identical to is itself. The most widely used expression of identity in the logical sense was explicitly formulated by Gottfried W. Leibniz in his Discourse on Metaphysics (Leibniz [1686], Leibniz [1714a]). Leibniz built upon what he called individual substance, which was to be the subject to which predicates were attributed but could not be attributed to another subject. Leibniz considered a person to be a substance since one can attribute predicates such as honesty to it, but not honesty. However, one could ascribe attributes to it since it can be attributed to something else. This is subscribed by Deutsch [1997], who proves that general similarity respects strict identity. Although strictly distinct objects may share the same property, strictly identical objects necessarily possess all the same properties. Objects, being ontologically distinct from properties, cannot themselves be properties, though they may instantiate them. Thus, strictly identical objects cannot differ in their properties, not because they fail to be properties, but because they are not properties at all. The distinction between identity and general similarity is a direct application of Leibniz’s Law. [...] all true predication is founded in the nature of things. When a proposition is not identical, that is, when the predicate is not explicitly included in the subject, it must be implicitly included in it. (Gottfried W. Leibniz in Discourse on Metaphysics, Section 8) The concept of separation is crucial in Aristotle’s Metaphysics: it exemplifies how two objects might hold the same form yet be distinct from each other (Berti [2019]). For instance, two chairs are the same in that they are chairs, but distinct in that they are two separate entities. This is echoed in Deutsch [1997]’s logical study of the sameness of shape between two objects —how they are indistinguishable with respect to certain properties— in proving that, though reduction to strict identity yields that general similarity is an equivalence relation, the indiscernibility properties of general similarity remain unaccounted for: to say that two objects are round and that they have a smooth boundary should be the same, but a smooth boundary is not a particular shape. He ultimately places the difference between identity and general similarity in the distinction between substance and attribute, again echoing Aristotle. 2.2. Relative identity 25 Lukasiewicz continues, the principle of identity is sterile, useless. 2.2 Relative identity Relative identity contends with two or more objects being identical in some relation but different in another relation, thus opposing absolute identity. It sprouted from an attempt at addressing the impracticality of the classical doctrine on identity when applied to the Catholic doctrine of the Holy Trinity, in which three entities are the same but different from each other —Father, Son and the Holy Spirit are all considered to be one God (Tuggy [2021], Van Inwagen and Howard-Snyder [1998]). It was first discussed by Peter Thomas Geach and Gertrude Elizabeth Margaret Anscombe in Three Philosophers (Geach and Anscombe [1950]) and further developed by Geach in The Virtues (Geach [1977]). It is not a particularly popular proposal. Despite a number of careful and creative expositions [. . . ], this notion of “relative identity” is widely viewed with suspicion. (Harry Deutsch in Identity and General Similarity) John Perry, for instance, assumes that any relation of identity, properly called, must satisfy an unrestricted indescribability schema —i.e., Leibniz’s Law. He uses this supposition in his arguments against relative identity (Deutsch [1997]). Although it remains controversial and has faced criticism from philosophers who argue that it may be unintuitive or potentially incoherent, Deutsch [1997] writes that relative identity is coherent and not inherently flawed. He suggests that relative identity can be understood as a logically sound concept, especially in contexts where objects may share certain properties while differing in others, which allows for a new perspective on change across time (§2.6) and contingent identity (§2.4). With this, he is responding to traditional objections, particularly those based on the necessity of Leibniz’s Law —which requires that if two objects are identical, they must share all properties (§2.1). Deutsch [1997] draws on the notion of general similarity (§1) as a way to formalize relationships where objects are similar but not strictly identical. In particular, he believes that certain entities, such as allographic objects (e.g., a work of art or musical composition), require a more flexible view of identity that 26 Chapter 2. Analytical Identity allows for multiple objects to be the same in different contexts or instances without violating the principles of identity. But how does relative identity defend such a definition of identity? At its core, it hinges on sortal concepts, a nomenclature coined by John Locke (Kaufman [2015]): countable terms (e.g., horse, person) through which objects and their properties may be classified. They can be divided into two types: phase and substance sortals. A phase sortal, such as a child or puppy, is a concept under which something can no longer fall without ceasing to exist —e.g., a child will eventually become an adult. Conversely, if something is to be considered a substance sortal, it must always be so: e.g., a rabbit at one time is the same rabbit at a different time, and that must be because the sortal rabbit applies to both. This, of course, is not completely hegemonic. Robinson and Weir [2024] follow Wiggin’s strong opposition of the concept, which led him to propose sortalism, a distinct idea for identity which hinges on sortals as well. Sortalism upholds traditional logic and Leibniz’s Law. Wiggins maintained that identity is absolute and that different procedures of individuation must yield the same answer regarding the coincidence of two objects (Noonan [2017]). Sortalism is based on the idea of substance and essential properties, while relative identity focuses on the relativity of identity statements. According to Noonan [2017], sortalism holds that identity statements always require specifying a sortal concept (e.g., same person, or same animal), while relative identity suggests they are always relative to a sortal concept. In order to identify an object, a sortalist must use a sortal term that answers same what?. Wiggins’ proposal has also received its fair share of criticism. Robinson and Weir [2024] present Michael R. Ayers’ criticism of his taking the notion of the body too loosely, because “he classifies as ‘lump-mass terms’ everything from bars of soap and pats of butter to pools of water and pots of stew” (Ayers [1974]). But let us go back to the relativist case. Any reflexive, transitive and symmetric relation may be seen as an equivalence relation. From there, one could infer that identity is an equivalence relation satisfying Leibniz’s law —but this is not always the case (e.g.,xand ybeing the same size or shape, as Deutsch [1997] exemplifies). However, Deutsch and Garbacz [2023] concede, in a sufficiently rich language, the difference between indiscernibility and identity is minimal to the point that the formulae for each coincide. A proof of it can be found in Epstein [2001]. The = symbol may be treated as a 2.2. Relative identity 27 constant for identity or represent a specific formula to be met by each set —which would make it language relative. Relative identity, like many of the proposals in the following sections, can be divided into a weak and a strong version. In its weak form, absolute identity is preserved and forced to coexist with true instances of relative identity (Deutsch and Garbacz [2023]). Advocates of this variant of relative identity hold that ordinary identity relations concerning the world of contingency and change are equivalence relations corresponding to restricted forms of Leibniz’s Law; and that the exact nature of the restriction depends on the relation. This allows absolutely different objects to have the same predicate but denies absolutely identical objects to have different ones (Deutsch [1997]). One of its defenders is Gupta [1980], who contends that identity comes in two forms: trivial (or absolute) and non-trivial (or relative). It tackles the Catholic Trinity (Tuggy [2021]) by proposing two predicates of relative identity for the creed —i.e., being the same person and being the same being (or God). Weak relative identity can deal with several paradoxes concerning identity. The Paradox of Change, which lies at the heart of the Ship of Theseus puzzle (§2.6, §2.8) may be one of the most notorious issues identity faces and one of the focal points of this work. We encounter it in everyday life with a simple question, most often bound to time: am I the same that I was? The changes an object undergoes throughout time violate Leibniz’s Law in that it ceases to share all properties. The main time-based solutiosn will be further discussed in §2.6. Deutsch and Garbacz [2023] explain that the weak version of ID considers an object in two separate times as the same object, but absolutely different things (i.e., different logical objects). The earlier instance is wholly present during its time and possesses the simple, non-relational property of not having the effects of ageing present in the older instance of that object. Deutsch [1997] presents us with the example of Fido the dog, which has seen some time pass and is now considered Fido the old dog. He proposes solving it on the basis that the identity expression is focused on one property (in this case, being a dog), and all other properties which are not subscripted to it should not be considered in thinking Fido the dog and Fido the old dog not indiscernible. 28 Chapter 2. Analytical Identity In doing this, he consents to falling at the hands of all philosophers who deal with the issue of object multiplicity, which warns of the infinite superposition of instances of the same object, due to the uncertainty of when one temporal part of it cedes its place to the next —meet Lucky and Patch in §2.6. To answer the Fido multitude (of temporal parts), he simply contends with considering only one of them (one of its proper parts) truly a dog and thus applicable to the identity relation at hand. According to Deutsch [1997], temporal parts and relations to time are not artefacts of extensional logic, while individuals are. Chrysippus, a Stoic philosopher from III BCE, sketched a puzzle on identity that would go on to resurface in Michael Burke’s Dion and Theon: An Essentialist Solution to an Ancient Puzzle (Burke [1994]). In it, a human called “Dion” is introduced alongside “Theon”, the proper part of Dion that includes everything but Dion’s left foot. After Dion loses his left foot, Chrysippus asks whether Dion or Theon survive after the amputation. If both survive the loss, Deutsch and Garbacz [2023] observe that two distinct physical objects occupy the exact space simultaneously. Assuming that is impossible and that, as commonsense demands, Dion survives the amputation, it follows that Theon does not survive. This conclusion is paradoxical because it appears that nothing happens to Theon in that interval that would cause it to perish (Deutsch and Garbacz [2023]). The puzzle involves two rigid terms, which do not see their reference (Dion and Theon) change with respect to parameters such as time (i.e., the moment in which Dion had both feet and when he no longer does) or possible world. This composes a necessary relation, which in turn links the puzzle to Necessary Identity and Necessary Distinctness (Deutsch and Garbacz [2023]). Necessary Identity holds that if two terms are equal, then the equation is necessarily true; conversely, Necessary Distinctness states that if two rigid terms are different, then the dissimilarity is necessarily true. Deutsch and Garbacz [2023] remark that the necessary truth of either relation does not imply the necessary existence of the referred objects: one might assume that what a rigid term denotes at a possible world (or moment of time) need not exist in it. A handful of solutions have been presented throughout the years for Dion and Theon’s situation. According to Philo of Alexandria, Chrysippus himself claimed that “Dion must remain while Theon has perished” (Long and Sedley [1987]). However, the opposite may also be arguable 2.2. Relative identity 29 —that Dion died and Theon took his place if it did not exist before (Deutsch and Garbacz [2023]). Then, three main categories of responses may be found. First, one can deny there is such thing as “Theon”, reasoning that non-separated proper parts of objects do not exist (Van Inwagen [1981]). Another option is to claim that neither Dion nor Theon exists, thus defending a reductionist view (e.g., arguing only elementary particles exist). A third proposal would entail embracing contingent identity (see §2.4) in affirming that objects of different kinds (a human and a proper part) can occupy the same space simultaneously. This might have been Locke’s answer to the problem, had he attempted to solve it (Kaufman [2015]). According to Deutsch and Garbacz [2023], weak relative identity adheres to the third category of responses in asserting that Dion and Theon survive the loss. At both times (before Dion loses his foot, and after) they are the same person, but they are distinct logical objects at first. This implies (by Necessary Distinctness) that Dion and Theon are distinct logical objects even after the amputation. Hence, one must allow that distinct logical objects may occupy the same space at the same time, which is unproblematic for this proposal: they are simply distinct logical objects that are physically coincident. A similar puzzle can be found in Searle [1980]’s Chinese Room argument, which was introduced in §1.2. In broad strokes, it presents a scenario in which Searle is alone in a room with instruction manuals, pen and paper; messages written in Chinese are slipped under the door, which Searle, who does not know Chinese, must answer according to the instruction manuals, and then slip his replies back out of the room. In it, Cole [1991] argued that a second individual may emerge from Searle’s activity as an ignorant translator —we will call them Polo, to honour Maloney [1987]’s harsh critique of the thought experiment. According to Maloney [1987], Polo shares Searle’s sensory and motor systems and goes wherever Searle goes, yet is distinguishable from him. In response, Cole [1991] rightly asks how can Polo be distinct from Searle and answers with psychological integration: the underlying unique psychology of an individual performing a task. Searle’s psychology will not be reflected in Polo’s performance, Cole [1991] argues because Searle has no access to the semantic content of the Chinese characters —even if he were to learn the manuals by heart. Searle [1980] had already noted that he could sit in an empty room, deprived of all manuals, having previously memorised every instruction, and, armed only with 30 Chapter 2. Analytical Identity pen and paper, write the outgoing strings of symbols —all the while he would still not know Chinese. Nigel Richards, the five-time English, two-time French and one-time Spanish champion of Scrabble, is a perfect example of this phenomenon: Richards speaks only English and has memorised the entire French and Spanish vocabulary without learning the language (Mitchell [2024]). Another paradox becomes relevant: the Paradox of the 101 Dalmatians, which presents the case in which thousands of proper parts could appear superimposed on the same dog. For each detail of the Dalmatian, even for each strand of fur, one would obtain another proper part of the dog. Applying this paradox to Dion and Theon, one finds that Theon is but one of the myriads of proper parts of Dion that result from the subtraction of every bit of Dion’s body, in every possible combination. This paper has already established that the weak version of relative identity is not particularly concerned with physically coincident but logically distinct objects (Deutsch and Garbacz [2023]). But the Paradox of 101 Dalmatians crowds a person with thousands of people. A maximal view of the issue would say that no proper part of an object is an object. Then, Theon is not a person. But Deutsch and Garbacz [2023] point out that issues quickly arise. They ask: when Dion speaks, do all the parts speak in unison? When he thinks, do all parts think the same? It becomes apparent that Theon cannot act without Dion, just as Maloney [1987]’s Polo could not act without Searle. There are two possible ways out of this junction: one must either deny that the many are people or that the people are many. Weak relative identity chooses neither: it does not support maximality. It rejects the idea that no proper part of a person is a person and supports considering all parts of people (Deutsch and Garbacz [2023]). Now consider a ball of clay (Clay) from which a statue (David) is made, which is destroyed the next day to make a second statue (Goliath). On the third day, part of Goliath is taken away and rebuilt with a different piece of clay (Claymore). This is a variation of Gibbard [1975]’s puzzle, and it will feature prominently throughout this thesis. One might think the relationship between the pieces of clay and the statues they form is identity, but Deutsch and Garbacz [2023] note that this conclusion contradicts Necessary Identity: if Clay is identical to David on the first day, then it follows that on the second day, David is Goliath (since Clay is identical to 2.2. Relative identity 31 Goliath on the second day) and therefore David exists then, which it does not. So one could either conclude that the constitution is not identity or that Necessary Identity is false. Gibbard [1975] original case argues for contingent identity (§2.4). Its responses range from arguing that Clay and David are not really identical since one existed before the other, and the latter can be destroyed by becoming a ball while the former cannot, to defending total coincidence as a plausible claim for contingent identity (sep-identity-relative). In terms of weak relative identity, contingent identity is regarded as absolute identity —the relation of David and Goliath to Clay is considered absolute. Since they are different statues, Deutsch and Garbacz [2023] reason, David and Goliath are different, but they are the same piece of clay (Clay), so this is a relation of identity. It follows that no name of Clay can be a rigid designation in the standard sense; only relatively, since it can always refer to the same piece of clay. The same is argued in Cole [1991] for machines. A physical machine has an intrinsic set of instructions through which it has the capacity to perform certain operations. This physical operation may store a program which can emulate a different machine —thus creating a virtual machine, a simulation of another intrinsic set of instructions. But Cole [1991] notes that those are different machines: they differ in properties. While the physical machine is running the emulation program, the virtual one may be running an application; they also present diverging intrinsic instruction sets and may have different speeds. He adds, even, that the virtual machine is dissimilar to the emulation program, as the program may be long or short, written in a language, contain comments or be copyrighted. In contrast, the virtual machine has none of these properties. In Aristotelian terms, Cole [1991] adds, a program could be but the form of some matter —it holds the power of an action, but without an underlying substance, it does nothing. According to the strong version of relative identity, there are (many) true instances of relative identity, but no such thing as absolute identity. It is difficult to defend, which has pushed most philosophers towards the weak version —moreover, Deutsch and Garbacz [2023] add, if taken literally, it is false. Geach defends this view (Geach [1991]). Geach’s arguments firmly rest on the imprecision of language, albeit changing in the course of his defence of strong relative identity. Geach [1967] contains the earlier, flawed argument that claims that the sameness of 32 Chapter 2. Analytical Identity truth conditions of a theory leads to the sameness of the entities whose predicates must be true if the said theory is to be true —i.e., ontological commitments. But Deutsch and Garbacz [2023] prove it wrong in noting that a theory is not ontologically committed to what must be in the universe for it to be true, but only to what must be true. Thus, there is no argument for the equivalence of truth conditions to the equivalence of ontological commitments. In Geach [1973], the latter argument, the relativity of language suggests that identity is indiscernibility, from which it follows that identical objects relative to one domain may be different relative to another. Geach claims that relative identity involves moving from a limited perspective to a more informed one (Deutsch and Garbacz [2023]). The relativity of language has led Geach, along with many others, to think that no domain is predetermined to be described. Still, theories shape the world, making things distinct or indistinguishable according to their descriptive resources. It thus relies on the Principle of Identity of Indiscernibles to say that indiscernibility implies identity. This might seem to go hand in hand with relative identity. But Deutsch and Garbacz [2023] warn that linguistic relativity is not really a kind of relativity of identity at all —if one point of view is able to detect a difference that another does not, and even if one is completely bound to the more restricted view, one must know that a difference exists. From Geach’s initial stance, others have taken the torch on adapting and defending relative identity. The proposal has faced many trials and rebuffs. One is the Passenger Problem, in which one person (Ava) may be counted twice as a passenger, so they would be two different passengers (Avi and Eva). If to say that Avi and Eva are the same person is to say that Avi and Eva are persons and are absolutely identical and to say that Avi and Eva are passengers and are absolutely distinct, a contradiction is found (Deutsch and Garbacz [2023]). One interpretation by Gupta [1980] tries to mend the problem by stating that the same person relation satisfies Leibniz’s Law. Still, the same passenger relation is not straightforwardly an equivalence relation. But this is incompatible with the Principle of Identity of Indiscernibles, Deutsch and Garbacz [2023] note, since if Avi and Eva are different passengers, then there must be some property that Avi has that Eva does not have. Although all the aforementioned paradoxes seem to be settled with relative identity, doubts 2.3. Composition as identity 33 quickly emerge. Most of its defenders claim that while identity relations are equivalence relations, they do not satisfy Leibniz’s Law in general. But the advocate of relative identity cannot simply reject any form of Leibniz’s Law. The problem is that of formulating and motivating restricted forms of Leibniz’s Law that are nonetheless strong enough to bear the burden of identity claims. Yet Deutsch and Garbacz [2023] point out that little systematic work has been done in this direction. Relativist theories of identity have received minimal support. There is not enough confidence that they really address numerical identity, can be internally consistent, and are sufficiently motivated. Proponents argue, however, that enough drive can be found in the many paradoxes and answers discussed in the literature. Moreover, no opponent of relative identity has provided evidence of inconsistency, and the weak view remains reliable by having a model within the theory of similarity relations (Deutsch and Garbacz [2023]). When an identity relation satisfies only a restricted form of Leibniz’s Law (i.e., preserving some properties but not all), one is left wondering how to determine which properties individuate the objects. Similarity relations themselves only conform to restricted forms of Leibniz’s Law, which is not a valid objection to require an identification of which properties are preserved. A property is generally kept by an equivalence relation if it spreads within an equivalence class determined by the relation —i.e. if one member of the class possesses the property, then all members do. This is what deutsch-similarity argues at the beginning of this section. 2.3 Composition as identity Composition as identity addresses the question of whether an identity can comprise other identities. It is motivated by the observation that there is a particularly close relationship between the whole and its parts (Noonan and Curtis [2022]). Nowadays, the discussion sees more criticism than defence: it is not a popular field, although it has seen a slight revival with Megan Wallace in the two-part series of articles Composition as Identity (Wallace [2011a] and Wallace [2011b]), along with Cotnoir [2013]. The main reason for its rejection lies in the steep price it seems to demand: by accepting that many parts are identical to one, we must give up Leibniz’s Law. There have been many attempts to fight against this grim conclusion, ranging from arguing for the discernibility of identicals in Baxter [1988a], to a more classical plea by Wallace to 34 Chapter 2. Analytical Identity keep Leibniz’s Law intact while advocating an amendment to first-order logic to include plural language and a hybrid singular/plural identity predicate (Wallace [2011a]). But let us not get ahead of ourselves. First, the landscape of the battle. The few defenders of composition as identity can be grouped into two main factions, one of which is in turn divided into two sides. Following the nomenclature in Wallace [2011a], we will call them the Weak Composition Thesis (WCT), the Strong Composition Thesis (SCT), and the Stronger Composition Thesis (RCT). They argue respectively that the composition relation is only analogously another form of the identity relation (WCT), that it is literally another form of identity (SCT), and that it is another form of identity, as well as saying that identity does not obey the Indiscernibility of Identicals (RCT) we introduced in §2.1. Baxter, for one, would fall under the RCT umbrella. Wallace defends SCT, while most of the other thinkers involved in this debate have preferred WCT. SCT takes the composition relation to be strictly the same as the identity relation —the parts are collectively identical to the whole. Although it would be grossly anachronistic to claim that Plato’s Parmenides (Plato) directly discusses composition as identity, since its modern formulation is more closely associated with contemporary analytic metaphysics, we may be able to find a first step towards SCT in it. The dialogue deals with the thought of the homologous founder of Eleaticism’s, who defended that the multiplicity of existing things, their changing forms and motion, are but an appearance of a single eternal reality —which he called Being (of Encyclopædia Britannica [2024]). In Plato, first Socrates and then Aristotle debated Parmenides on the nature of ideas. While discussing transcendental questions such as existence and the one and the many, the subject of sameness is raised: Parmenides: If one is, the one cannot be many? Aristotle: Impossible. Parmenides: Then the one cannot have parts, and cannot be a whole? Aristotle: Why not? Parmenides: Because every part is part of a whole; is it not? Aristotle: Yes. 2.5. Identity across worlds 41 2.5 Identity across worlds One of the central debates in identity theory concerns whether identity can be a continuous concept across different worlds of possibility, each representing a certain way in which objects might have been arranged or have had different characteristics. This topic got considerable attention in the 1960s and 1970s, largely due to developments in modal logic. Two main theories (though by no means the only ones) have been debated extensively up to the present day: counterpart theory and transworld identity theory. Transworld identity defends that identity might spread throughout different worlds while keeping its defining properties —Mackie and Jago [2022] clarify that it should more appropriately be referred to as interworld identity or even transmodal identity. Its advocates can be roughly divided into two schools: one seeks to narrow down what are the essential properties of an identity, rather than its accidental or changeable ones (i.e., contingent); the other attempts to split an object into parts and then examine how many of those can change before it loses its original identity. Both contrast with haecceitism, which maintains that every individual entity possesses a unique, intrinsic quality (haecceity or thisness) that makes it, unlike any other entity, regardless of its properties or relations to other things (Mackie and Jago [2022]). The basic tenet of transworld theory is that an entity can exist in more than one possible world (Noonan and Curtis [2022]). When modal discourse is viewed through this lens, it becomes natural to interpret statements attributing modal properties to individuals as asserting their identity across those worlds (Menzel [2024]). Modal statements about what might or could have been the case, what could not have been otherwise, and so on, are often characterised within the possible-worlds framework. Mackie and Jago [2022] exemplify this with the expression Bertrand Russell might have been a playwright instead of a philosopher, which is interpreted as “there is some possible world in which Bertrand Russell exists and is a playwright rather than a philosopher”. According to Lewis [1986], there is no objective difference in ontological status between the actual world and other possible ones. We refer to our world as actual simply because we inhabit it; the inhabitants of another world may, with equal rights, regard their world as actual. In other words, actual in the actual world is but an indexical term (like here or now). Mackie and Jago [2022] go on to extend the possible-worlds interpretation to de re 42 Chapter 2. Analytical Identity modal statements, which involves a commitment to the idea that some individuals exist in more than one possible world, leading to the concept of identity across possible worlds, or transworld identity. A common example of this is the utterance Sarah might have been a millionaire, which can be understood as asserting that there is a possible world in which an individual identical to Sarah is a millionaire. Conversely, Sarah could not have been a millionaire implies that in every possible world in which an individual identical to Sarah exists, that individual is not a millionaire (Noonan and Curtis [2022]). The difference between de re and de dicto modal utterances becomes central to this discussion. A notable problem presents itself when considering the identity of complex artefacts across possible worlds. Noonan and Curtis [2022] exemplify this with a bicycle, for which it seems reasonable to assume it could have been made of a number of different parts. However, it does not follow that the same bicycle could have been constructed from a completely different set of parts. In a sequence of possible worlds starting with ours, each containing a bicycle slightly different from the one in the previous world, the last world might contain a bicycle made entirely from different parts to the one in ours. It would be incorrect to say that each bicycle is identical to the one in the neighbouring world but not identical to the one in a distant world, given that identity is a transitive relation. The notion of the same object existing in more than one possible world, with the actual world being treated as one of them, is a central issue in the debate about identity across worlds (Noonan and Curtis [2022]). Opinions range from deeming identity between objects in different possible worlds so problematic that it is unacceptable to defending it as completely uncontroversial. This debate is further complicated by the introduction counterpart theory, which replaces the claim that an individual exists in more than one possible world with the idea that each individual exists in only one world —but has counterparts in other worlds. Counterparthood, based on similarity, does not follow the logic of identity. Going back to the bicycle example, Noonan and Curtis [2022] concedes that since similarity is not transitive, it allows one to argue the bicycle could have had some different parts without implying that it might have been totally different. This scenario leads to the concept of counterparts, as developed by Lewis over more than twenty years in Lewis [1971], Lewis [1986] and 2.5. Identity across worlds 43 Lewis [1993]. At its core, the theory argues that an individual does not exist in more than one world but has counterparts in other possible worlds —entities that resemble them in relevant ways. This is, Mackie and Jago [2022] write, the alternative to transworld identity: instead of saying that an object (in whole or in part) inhabits more than one world, counterpart theory, with Lewis as its maximum representative, holds that it belongs to one world (the actual one) and has counterparts in the others. I t is those counterparts who experience the aforementioned changes in properties. But let us go back to transworld identity. The theory revolves around the concept of transworld properties, which would serve as a sort of anchor to fix an individual’s identity as one and the same, albeit existing in more than one world (Orilia and Paolini Paoletti [2022]). They raise important questions, Mackie and Jago [2022] note. For example, do properties themselves have essential characteristics that allow them to be identified across worlds? Is the identity of a property across worlds tied to its causal or nomic role9? Alternatively, do properties have bare identities? Causal essentialism posits there is nothing to the identity of a property beyond its causal role. Defenders of this proposal argue that a property’s essence is defined by its causal and nomic relations (Gibbs [2018]), while others follow Fine [1994] in understanding essence as a more fundamental, primitive notion than modality, thus reversing their priorities. In contrast, Mackie and Jago [2022] add, quidditism claims that properties have bare identities, thus becoming the haecceitism of property identity. Among other authors, Lewis [1986] goes on to defend a version of it, although it is not free from criticism. Relational modal predication considers properties as relative to the world in which they exist. Another problem in the debate on transworld identity is whether different objects are merely added together or paradoxically superimposed. Mackie and Jago [2022] invite us to consider this: if a cow has four legs but could have had three, does it have seven legs when one adds together both possible worlds, or does it have four legs and three legs at the same time? Proponents of extreme realism might argue that the cow has a four-legged part in our world and a three-legged part in some other world —thus adhering to counterpart theory. Relational modal predication posits that the cow has four legs in one world and three in another. Its advocates argue that an 9The lawful or law-like relationships and behaviours a property or entity exhibits within a system of natural laws. 44 Chapter 2. Analytical Identity individual can have both four legs and three legs without contradiction, provided the predications are relative to different possible worlds. Yagisawa [2010], following in Lewis’ footsteps, extends this idea to suggest that concrete entities such as trees, cats and cows are extended across worlds by having stages (or parts) that exist in each of them. Known as stage theory, this notion will resurface throughout this thesis, starting with the following section on identity through time (§2.6). According to this view, ordinary entities consist of spatial, temporal and modal stages, all of which are equally real and metaphysically equivalent to temporal and spatial stages. This is known as Four Dimensionalism, which we briefly saw in §2.3. Furthermore, it is crucial to distinguish between essential and accidental properties, a terminology that has prevailed at least since Aristotle (Mackie and Jago [2022]). Essential properties are those an object must have in order to be the object it is, while accidental ones are those it can lack. For instance, the cow’s species, biological structure, and genetic code may be essential properties, while its number of legs is accidental. This distinction is important when considering identity across possible worlds since an object’s identity can be preserved across worlds if it retains its essential properties, even if its accidental properties vary. Yet Mackie and Jago [2022] warn that there is no consensus on which ones are truly essential, and whether properties are fixed or can vary with context or perspective. Finally, the debate between endurance and perdurance theories of identity over time (§2.6) also informs discussions of transworld identity. Endurantists believe objects are fully present at every moment of their existence, while perdurantists argue they persist by having different temporal parts at different times. This parallels the debate over whether objects exist in more than one possible world (endurantism) or whether they have different modal parts in different possible worlds (perdurantism). The more than striking similarity of these two debates will continue to manifest itself in the following section. 2.6 Identity across time The problem of identity through time finds its greatest enemy in change. Accepting the assumption that there is change seems inconsistent with embracing Leibniz’s Law, which states 2.6. Identity across time 45 that a thing cannot be one and the same before and after the change. Strictly speaking, then nothing would have really changed: one thing would wholly become another. This reflection is also applicable to Artificial Intelligence. In the case of a Neural Network, the underlying system that adjusts the weights and probabilities of the model is constantly changing as it learns (Cole [1991]). Numerous thinkers have taken up this challenge with varying degrees of success. The discussion even preceded Aristotle, who distinguished between accidental (e.g., a car whose tyres have been replaced, thus preserving its identity even after the change) and essential (e.g., a car that has been turned into scrap metal in a junkyard, without preserving its identity) changes (Barnes [1977]). According to Aristotle, a thing can be one and the same after undergoing accidental changes. Aristotle [2018] himself criticises some of his predecessors’ proposals on identity across time: And further, observing that all this indeterminate substance is in motion, and that no true predication can be made of that which changes, they supposed that it is impossible to make any true statement about that which is in all ways and entirely changeable. For it was from this supposition that there blossomed forth the most extreme view of those which we have mentioned, that of the professed followers of Heraclitus, and such as Cratylus held, who ended by thinking that one need not say anything, and only moved his finger; and who criticized Heraclitus for saying that one cannot enter the same river twice, for he himself held that it cannot be done even once. (Aristotle in Metaphysics, Book Gamma, 1010a 7 - 1010a 16) Many have followed him, a small group of whom we will meet throughout this section. The discourse on identity across time translates the worlds problem to time itself, pondering whether an identity can withstand the passage of time. The present can be thought of as the actual world, while both the past and the future are possible worlds. Although rich in problems of its own, it is perhaps most recognisable by one of the most famous riddles of identity: first 46 Chapter 2. Analytical Identity introduced by Plutarch in Chapter XXIII, Section 1 of Life of Theseus (Plutarch [1914]) and expanded by Thomas Hobbes in Of Identity and Difference (Hobbes [1656]), it follows the life cycle of the ship of Theseus as it is dismantled and rebuilt. The Ship of Theseus is a symmetrical fission case. Identity across time problems can be classified in a roughly binary fashion and thought of as being either about fission (i.e., splitting an object) or fusion (i.e., joining objects). In symmetrical cases, there is some relation that each of band cstand in to a, where standing in that relation is what makes each one theoretically identical with a. Conversely, in asymmetrical cases, there is a certain relation that makes btheoretically identical to a, while cis theoretically identical to abecause it stands in a different relation to a(Gallois [2016]). As with the worlds problem, identity across time hosts a myriad of theses. Braddon-Mitchell and Miller [2020] list three standard theories of persistence to categorise them: endurantism, perdurantism and exdurantism, which argue respectively for continuous numerical identity, Four Dimensionalism (or worm theory) and object-stages. In perdurantism and exdurantism, which will sometimes appear merged under the umbrella of perdurance theory, an object persists in time by having temporal parts of the same object existing at different times (Gallois [2016]). The changing properties we might observe on an object would each be assigned to different temporal parts. Among other proposed variations, stage theory, proposed by Hawley [2001] and Sider [2001] may help better understand perdurance. Stage theory proposes a change in semantics, now considering each temporal part of an object an instance of that object. Gallois [2016] warns that one must be careful to distinguish whether one is counting based on absolute identity or not when adhering to this theory —i.e., counting objects across time (e.g. someone who says they have had three cats in their life) or object-stages (by which those three cats could be one and the same cat). Christopher Conn argues that four-dimensionalist ontology best captures what might be going on in Locke’s thinking about persistence and the relationship between persons (or organisms) and finite intelligences. Organisms, Kaufman [2015] writes, according to Four Dimensionalism, are persistent rather than enduring entities —a distinction whose locus classicus is Lewis [1986]. If worm-theoretical perdurantism is correct, then the same person relation holds between person-stages by virtue of those stages being parts of the same person. If stage theory 2.6. Identity across time 47 (i.e., exdurantism) is correct, then the same person relation holds because one person-stage is a temporal person-counterpart of the other (Braddon-Mitchell and Miller [2020]). Endurance theory states that an object has spatial parts that are fully present at the different times it exists, rather than temporal ones (Gallois [2016]). The object endures and changes through time but only extends through space —thus, these must be sharply distinguished from events or processes which do extend through time. Objects have different (modal) properties from the summations of object-states with which perdurance theory identifies them. §2.5 introduced the concept of modal property. Considering the close relationship between the debates on identity across worlds and identity across time, it is of no surprise that modal properties should be featured in this section as well. Just as a person might be taller in another world, so a child might be taller in a few years. The issue is that, in time, a process of change is necessary, yet we may see otherworldly alterations as a simple, somewhat static alternative. A crucial distinction is made between identity at a time and identity across time (Gallois [2016]). Synchronic identity refers to identity at a single point in time: the book I am reading is (now) identical to the one next to me. On the other hand, diachronic identity refers to identity holding between something existing at one time and something existing at another: the book next to me is identical to the one I acquired six months ago. Geach addressed both synchronic and diachronic identity by denying the existence of a single absolute relation of identity, instead positing a variety of relative relations of identity. We have reviewed his proposal in §2.2, which involves the sortal concepts Fand Gunder which aand bfall; amay be the same Fas b, but not the same Gas b(Deutsch and Garbacz [2023]). Continuing with the examples in this paragraph, if the book I am reading is identical to the one next to me, then there must be a substance sortal (the concept of a book) in both the book I am reading and the book next to me share. Moreover, Deutsch and Garbacz [2023] add, a substance sortal is said to accompany a criterion of identity —which is, among other things, a criterion for some earlier Sbeing the same Sas some later Sassociated with substance sortal S. One way the relativity of the answer to an identity question can help deal with the shortcomings of the classical approach to diachronic identity takes us back to the Paradox of the 101 Dalmatians, which we examined when we objected to relative identity in §2.2. In it, we found 48 Chapter 2. Analytical Identity a Dalmatian, which might have had as many proper parts as there were strands in its fur, each of which encapsulated the whole dog but one hair. We will simplify this example so that only two entities are observed: the Dalmatian (which we will call Lucky) and a proper part (Patch). Deutsch and Garbacz [2023] ask us to suppose that nothing is a dog if it is a proper part of a dog. Then Patch does not qualify as a dog when it is a proper part of Lucky. In this case, a relative identity theorist might say that at a later time, Lucky is the same dog as Patch. Since Patch was not a dog at first, neither Lucky nor Patch was the same dog at a later time as Patch at the beginning. So, it seems that identifying Lucky with Patch as the same dog later on does not require identifying Lucky with one of its proper parts, Patch, at first. Criteria of identity can thus be applied synchronically to determine whether two coexisting objects are parts of the same object of a sort or diachronically to determine identity over time. The same entity may exist at more than one time, which means it is possible for a thing existing at one time and a thing existing at another time to be numerically identical (Noonan and Curtis [2022]). Heraclitus famously argued (or at least others wrote that he did) in Plato’s dialogue Cratylus (Sedley [2003]) that one could not bathe twice in the same river because new water was always flowing, and neither could the same one. This section has featured Aristotle criticising the notion explicitly (Aristotle [2018]). David Hume, on the other hand, argued that identity over time is a fiction we replace with a collection of related objects (Speaks [2006]). But this brings us back to the apparent inconsistency with Leibniz’s Law: a person at different times will not share, at least, the property of age. 2.7 Vague identity The impossibility of vague identity, similar to the impossibility of contingent identity, seems to follow directly from the classical concept of identity (Noonan and Curtis [2022]). It argues that the identity relation between objects need not explicitly match, thus rejecting the absoluteness of identity —similarly to what relative identity did in §2.2. Such relation can be found when one or both terms flanking the identity relation are imprecise designators (i.e., indeterminate or ambiguous referents). Its impossibility is easily illustrated by Noonan and Curtis [2022]’s example involving a, which is only vaguely identical to b. Then something is true of a—i.e., 2.7. Vague identity 49 that it is only vaguely identical to b— which is not true of b. So, by Leibniz’s Law, ais not identical to bat all. So, it would seem that this kind of vagueness is only possible, at most, if one or both of the terms flanking the identity sign are imprecise designators10. However, it could be argued that this is too blunt. Even identity statements with precise designators may be indeterminate in some cases. To take a well-known example, consider Mount Everest and a well-defined snowy geological formation —which we will call Rock. Whether the boundaries of Rock coincide with those of Everest is indeterminate, so, as tempting as it is to think that both are precise designations, the identity statement Everest is Rock remains indeterminate (Noonan and Curtis [2022]). Yet the proposal has known less defenders than detractors. Terence Parsons’ book Indeterminate identity: metaphysics and semantics (Parsons [2000]) provides the most thorough defence of the theory while, most notably, Evans [1978] and Salmon [2005] object to it on the grounds that it violates Leibniz’s Law —as seen in aand b’s example. This debate has been going on for decades and has involved names we have become quite acquainted with in this chapter, such as Lewis and Van Inwagen. Hawley [2001] presents a variant of Evans [1978]’s argument on the impossibility of vague identity. Hawley [2001] presents an example involving Van Inwagen [1990]’s fiendish cabinet, which disrupts all features relevant to personal identity (Noonan and Curtis [2022]). In Hawley [2001]’s example, an individual called Alpha enters the cabinet, and another individual, Omega, eventually exits. It is indeterminate whether Alpha steps out of the cabinet, and Alpha is such that it is indeterminate whether she steps out of the cabinet. Conversely, it is not indeterminate whether Omega steps out of the cabinet and Omega is not such that it is indeterminate whether she steps out of the cabinet. Hence, Hawley [2001] concludes that Alpha is not identical to Omega. Evans [1978] had located the indeterminacy in language, finding it too imprecise to convey what was needed to articulate identity unerringly. Noonan and Curtis [2022] illustrate the validity of this conclusion in cases of synchronic identity (§2.6) with a structure which consists of two halls: Beta and Psi, connected by a scanty corridor. Alpha is in Beta, while Omega is in Psi. The nature of the structure makes it indeterminate whether the identity statement The 10A term or expression that does not uniquely or precisely pick out a particular object or entity. 50 Chapter 2. Analytical Identity building in which Alpha is located is the building in which Omega is located is true or false, since it is unclear whether Beta and Psi are two separate buildings or merely parts of a single one. Shoemaker and Swinburne [1984] agree in that the term building is too vague in this case to settle the issue. In Hawley [2001]’s example, one could argue that multiple entities are present: one that exists before and after whatever happens in the cabinet, one that exists only before and one that exists only after. This would be consistent with what an adept of perdurance like Quine [1960] would say, who would be willing to treat any spatio-temporally disconnected regions as containing a physical object. On the other hand, a persistence theorist may choose to accept the problem as one of genuine vagueness, rejecting the multiplicity of entities posited by perdurance and instead discarding Evans [1978]’s argument against vague identity (Noonan and Curtis [2022]). This position is more closely associated with common-sense views of identity. 2.8 The Fleet of Theseus The Ship of Theseus is one of the most famous identity puzzles in History. Not only has it been discussed extensively by many philosophers over the years, but it is often used to illustrate the difficulty of defining identity in everyday life. It has appeared in pop culture time and again. It is a symmetrical fission problem relating to identity across time (Noonan and Curtis [2022]). Its main focus is to test it through the application of the Paradox of Change (§2.2), which Noonan and Curtis [2022] define as stating that an object that changes parts violates Leibniz’s Law on the grounds that it no longer retains all of its original properties and is thus no longer identical to what it was. Although it represents only a tiny facet of the challenges of identity, it will serve as an illustrative comparison between the myriad of answers humanity has come up with to one question: where does the ship of Theseus’ identity lie? The setting takes us to a new ship, made entirely of wooden planks, called the Ariadne. The Ariadne sets sail, and as it traverses the seas, its planks are gradually replaced, one by one, by new ones that are descriptively identical to the original ones. The old planks are taken ashore and stored in the port of ancient Athens, Piraeus. After all the planks have been replaced, the 57 With this crash course in Locke’s theory of identity, let us now broaden our horizons. We move forward close to a century. Hume admits to having no clue what a soul might be other than a bundle of perceptions, which are clearly changeable and bound to a mortal body. In On the Immortality of the Soul (Hume [1777]), he denies us having any idea of the self since it is never separate from perception. Thus, Speaks [2006] remarks, we never know a self beyond the mental events that compose our mind in a chain governed by cause-effect. Hansen [2016] points out that, curiously, this remark is very close to what Vasubandhu, an Indian philosopher and Buddhist monk, advocated in the c. 5th century: that the self is a continuum of aggregates, which are the physiological elements which make up the individual person. Hume concludes, rather brutally, that the subject of personal identity can never be decided. It may seem odd to include him in the field of personal identity, but he very much discusses it. According to Speaks [2006], Hume’s central idea is this: we confuse the ideas of identity and diversity (or distinctness). He attacks the age-old issue of change over time, particularly at a gradual pace: we tend to mistake qualitative (or, as he puts it, specific) identity and numerical identity (which is the one that is lost in change) —a distinction we saw when introducing Leibniz’s Law in §2.1. And to try and fix our mistake, we invent some aspect (the mind, soul, self, or substance) which has somehow remained unchanged throughout, thus serving as an anchor of sorts for the object’s identity. He renounces this fix and modulates objects not to be identical throughout time, but merely related. The same, he says, applies to people. Hume’s theory of personal identity, Robinson and Beauchamp [1978] assert, therefore, holds that the self is entirely reducible to a collection of perceptions and that it is not characterised by strict identity in the sense of sameness of substance. There is nothing simple in the collection of perceptions, and both identity and simplicity are ascribed to the self because of the causal relations and similarities between the perceptions which unite them. Conversely, there are no real connections or relations between these perceptions that unite them (Robinson and Beauchamp [1978]). Although we saw similar conclusions in Locke, it is notable, Speaks [2006] points out, that the latter does not seem to be fazed by objects being numerically identical to change. He also sets memory (or extended consciousness) as a key ingredient for personal identity, which Hume does not —the causal relations between mental events are what makes memory possible, 58 Chapter 3. Personal Identity which he considers too fragile for keeping as important a link between a person and his future, maybe more forgetful, self (Speaks [2006]). Thus, the sameness of persons is explained in terms of psychological characteristics and relations, rather than material identity or persistence of an immaterial soul. Vasubandhu, on the other hand, believed that one’s memory of an object is awakened when a special function of the mind connects with earlier occurrences and identifies its object (Hansen [2016]). Both Hume’s and Locke’s views enjoy a contemporary following that still defends variations of their arguments with unwavering conviction. But they are by no means the only ones on the field. Joseph Butler, an English Anglican bishop, is known for his critique of Locke’s theory of personal identity. In Of Personal Identity (Butler [1736]) he argues against Locke’s view that personal identity is constituted by consciousness or memory. His main criticism is that Locke’s theory is circular: he holds that consciousness of past actions presupposes, rather than constitutes, personal identity. Butler [1736] maintains that Locke made a wonderful mistake in failing to recognise that the relation of consciousness to personal identity is one of evidence, not constitution. Butler [1736] does not focus on the continuity of consciousness in the way that Locke does, nor does he explicitly advocate a Cartesian dualist conception of the soul. Instead, he emphasises the logical problems with Locke’s account, claiming that we have an intuitive understanding of our own identity that goes beyond mere consciousness of our past actions. He suggests that personal identity is something simple and indivisible that cannot be reduced to or constituted by a series of conscious experiences. Furthermore, Butler [1736] argues that Locke’s theory fails to account for the unity of consciousness over time. He contends that if personal identity consisted of consciousness alone, as Locke suggests, we would have a new person with each new conscious thought. This, Butler [1736] holds, goes against our intuitive understanding of personal identity as something that persists across time. On another side of the galaxy, Jacob Berger wrote A Dilemma for the Soul Theory of Personal Identity (Berger [2016]). He strayed a little farther from Locke and into the system of soul theory, to which Butler adhered. Soul theory explains, according to Berger [2016], the persistence of personal identity with non-physical souls that appear to occupy the space left by the mind. An example of a soul in the sense he discusses would be the Hindu concept of ¯ Atman: an 59 eternal spiritual core of a person, independent of the body and particularised interests —it is also close to the Platonic soul. Among other proposals, Richard Granville Swinburne’s idea of indivisible, unchanging souls that may take on different properties over time not only echoes Plato’s Phaedo’s (Church [1954]) immutable and metaphysically simple souls (without parts): it mirrors the idea we saw with Clay, David and Goliath (§2.2): in this case, the soul is always Clay, simply taking on different shapes. Berger [2016] contends with it through a dilemma: if souls change over time, then the central question of personal identity remains unsolved. If they remain impervious to time, then soul theory can, at best, deal only partially with personal identity since it would be unable to explain why souls relate to particular bodies over time. He concludes that it is no better than bodily or psychological theories of continuity, such as those of Butler or Locke. A thought experiment by Locke [1689] to illustrate that the body (and the soul) is not the proper criterion of personal identity may be worth contrasting here. He presented us with a prince and a cobbler whose minds radically change until the prince’s body behaves like the cobbler and vice versa. A common intuition is that the prince is now inhabiting the cobbler’s body, Berger [2016] writes —lest we think the cobbler has simply gone mad. On top of that, if the soul of the prince remains in the prince’s body, as one might assume, then the soul theory would seem to deliver the wrong result —thus, Berger [2016] notes, the soul is not a proper criterion either. This is supported by Leibniz’s King of China example, in which most would refuse to transfer their soul to another body, which would keep all its physical and psychological capabilities. Shoemaker explains it by saying that our practical concerns typically track our psychological or bodily features, not whatever soul features we may have. But our interstellar travels are far from over —let us now move on to another constellation of issues. In §2.3 we encountered the problem of the many, for which we took the words of Baxter (Baxter [1988a]) and Wallace (Wallace [2011a], Wallace [2011b]). In Material Beings (Van Inwagen [1990]), cited as one of the most influential works on metaphysics of the late 20th century, Van Inwagen restricts the special composition question —when some individuals compose a complex object— to life or organisms. In doing so, he avoids the usual identity pitfalls, such as the Ship of Theseus (§2.8). The solution is simple: ships do not exist, so there 60 Chapter 3. Personal Identity is no problem. The issue is that other cases of organismal fission have been formulated, such as brain-duplication scenarios (see §3.1). His reasoning has been challenged by other theorists such as Eklund [2002], who finds unconvincing his almost aggressive defensive strut about being the only non-arbitrary moderate solution —as opposed to radically considering that no simples ever compose complex objects, or that any simple could. Moreover, while he focuses on justifying organisms exist, she poses he should be concerned with why, if no other complex objects exist, organisms do. Van Inwagen’s solution to the problem of the many is fuzziness. He claims that it is an ontologically vague matter which individuals make up a given life —one is composed of a fuzzy set of simples, membership of which is a matter of degree. Eklund [2002] promptly knocks it down with a sweep: there are many slightly different fuzzy sets, but Van Inwagen does not get us around the absurd conclusion that there are many lives where we thought there was only one. Problems with his refusal to accept contingent identity (§2.4) promptly follow. We are now approaching the first destination of our travels in this chapter —a star known as Brain Duplication (§3.1). Ehring [1987] discusses the explanations offered by the orbiting planets we will visit in §3.2 and §3.3: mentalistic accounts of personal identity and, more specifically, memory-based arguments for psychological continuity. As with most things in these discussions, intuition causes our headaches. After all, if we split a brain in two and somehow manage to implant one half into a corresponding body, which of the two resulting entities (which share all past memories) is now the original being? If they both are, that would mean they are also each other, which is intuitively absurd. Ehring [1987] sets out to bring this meddling intuition to heel. Usually, he writes, the evidence that points in the direction of believing there are two people after the procedure is the lack of a shared consciousness and a common body and the causal independence of the duplicates. On the assumption that one person can possess only one mind at a time, we conclude that the two new entities are two persons. The same goes for bodies. Yet Ehring [1987] proposes a scenario that aims at showing these conditions are not sufficient. We are presented with John, a man in his thirties who hops in a time machine and goes back in time to hold a conversation with his four-year-old self: is John talking to himself? Ehring [1987] says yes —moreover, he sees it as the most natural description of the scene. He avoids difficult observations, such as the incompatibility of their characteristics, by considering them 61 stages of a person, thus adhering to perdurantism. But that leaves us with the third piece of evidence: physical-causal independence between duplicates. The time travel example does not cut it, given the causal relation between both stages (whatever happens to four-year-old John will be reflected on thirty-something-year-old John) —but we find that in brain-duplication cases, once they are nestled in their respective bodies, the duplicates are independent. To address it, Ehring [1987] finally goes to the memory theorists, with whom we will become more acquainted in §3.3, and claims that, through their shared thread of memories (albeit it eventually bifurcates), the split brains can be considered stages. An alternative view can be found in Prior [1957], which turns us into an amoeba. He presents himself and the reader as being the two resulting amoebas after asexual reproduction. Aside from the no doubt complicated nature of this scenario’s interpretation, Prior [1957] has no issue with simply considering each amoeba a distinct individual, with the particularity that their lives can be drawn in a Yshape that shares a pedestal during which they both were themselves and each other. Humans are, almost from birth, imitative creatures. (Tony E. Jackson in Imitative Identity, Imitative Art, and AI: Artificial Intelligence) In the field of Artificial Intelligence, one could approach this discussion from the perspective of imitation, which Jackson [2017] describes as a purposeful creation that looks or sounds like something else that is not an imitation. Turing [1950] himself used the term in describing what would come to be known as Turing’s Test: his imitation game. Imitation is so pervasive in humans and animals alike, Jackson [2017] writes, that even Aristotle inferred that it is something like an instinct to our nature: we imitate facial expressions, voices, movements, and emotional states. Our disposition to imitate is so strong, so automatic or unconscious, that not doing it is the exception, and it becomes an unconscious constituent of our conscious identities. Human identity is, even on the level of neurobiology, an imitative identity. The history of much of Western art has barrelled toward photographic realism, which Jackson [2017] thinks is related to our innate fascination with stories involving what he calls over-imitation. 62 Chapter 3. Personal Identity So, it should come as no surprise that we ended up dreaming of creating something that would imitate us perfectly, even at the risk of misbelief —a peril that Jackson [2017] links with its enjoyment. This is the same reason we like to watch magic shows. Humanity began by sculpting realistic humans, kept the dream alive with animatronics, and is now reaching for the promise of Artificial Intelligence. But this pursuit comes with a threat. Jackson [2017] warns of the possibility of being so successful that we cannot tell it is an imitation. This is a profound danger, he writes, not just at the level of psychology, but at the level of ontology: the status of our being as entities in the world. These are, of course, just a few of the many proposals on this topic. We will explore additional perspectives in the following sections. To conclude the introduction to this chapter, I draw upon the characters from John Perry’s A Dialogue on Personal Identity and Immortality (Perry [1977]): Gretchen Weirlob, a fatally wounded philosophy professor, engages in conversation with Sam Miller, a chaplain and friend, and Dave Cohen, one of her former students. Weirlob, a seasoned realist, both encourages and entertains Miller’s attempts to persuade her of the possibility of an afterlife. Their discussion will serve as a summation and final expansion of the ideas in these pages as we prepare to dive into the first section of the chapter. The first two nights focus on the philosopher and the chaplain, while the third features a specific case of a brain transplant discussed by all three characters. It is not until the end of the second night that Cohen takes the floor. Weirlob gives yet another definition of personal identity as the dialogue takes flight: it is a relationship that encompasses both the anticipation of a future self who will continue to experience life and the memory that preserves and accounts for the person one was in the past. This may ring a Lockean bell if we understand this temporal stretching as a consciousness reaching for both the future and the past. It will also come back in the next chapter, on Cultural Identity (§4). Moreover, as we saw with Deutsch [1997] in §2, a distinction is made between identical in the sense of similarity (however exact) and identical in the sense of sharing an identity. A person being more than a body is then discussed —and amongst the soul, self, or mind, consciousness arises, as if Locke were truly being resurrected. His same argument is used by Weirlob (a.k.a, Perry) on the insufficiency of the soul (Cartesian immaterial substance) in matters of identity. 63 Without certainty about the correlation between a body and a soul (here comes the dilemma we saw with Berger [2016]), it cannot be established as fact. This seems to have a strong flavour of experimentalism1. Though it is supported by the subjective experience of having the same identity (and soul), which is then generalised, the issues do not seem to be addressed. Nor is an observable link established between body and soul: the similarity of psychological characteristics (a person’s attributes, beliefs, memories, prejudices), the responsibility for which is laid on the soul or mind. But just as a river’s water can be different while its similar states are enough to identify the same river, the different states of mind do not require the identity of the mind. So the objection still stands: whether they exist or not, Perry [1977] writes Weirlob, clinching her argument that immaterial souls are irrelevant to personal identity. To further complicate things, composition as identity (see §2.3) is brought to the mix: identity statements of an object of a certain kind involve judgments as to their parts being connected in a certain way. So, the identity of persons is not just the body, the soul or the mind —it is a whole composed of stretches of person-stages, of consciousness. With this, we have moved to the territory of Four Dimensionalism (§2.5), which was somewhat preceded by Locke’s stretching consciousness. In the end, Perry [1977]’s characters find themselves lost in the modern issue of circularity —which we will further discuss in §3.1. This leads to Cohen (who has finally arrived) defending that personal identity can be explained in terms of memory, which we will discuss further in §3.3. Memory, however, is not sufficient if duplicates cannot be identical to the original person. And even if one were to argue that it would be sufficient as long as there was only one continuation, then identity would be based on uniqueness, which is extrinsic to the person. We will save Perry [1977]’s third night for §3.1. But before we go any further, here is a disclaimer. This chapter presents many ideas and examples that we might even find absolutely preposterous, in order to prove a promising theory wrong. This is a product of the extremely high expectations philosophers have of the concept of personal identity: it is not that they want it to fit our current understanding and possible edge cases of identity, but that they want it to be a universal concept. Universality implies 1Experimental philosophy, which emphasised empirical observation and controlled experiments, was gaining ground as an approach to natural science in Locke’s time. Locke himself was an advocate of this approach, though his views on the matter had their own distinctions. For example, he emphasised sensory experience as the source of all ideas (Gordon-Roth [2020]). 64 Chapter 3. Personal Identity blanket applicability, whether the recipient is a Martian, a time traveller, a clone, or half a brain in an alien body. As such, an argument that might veer into many-worlds discourse (§2.5) or straight-up science fiction will be commonplace, perhaps even difficult to follow at times, and not to be taken lightly. 3.1 Brain duplication against personal identity Perry [1977] presents us with a difficult scenario on the third night, taken from Barbara Harris’ 1977 book Who is Julia?. Julia North and Mary Frances Beaudine are both transported to the hospital in opposite yet equally fatal conditions, and Julia’s healthy brain is taken out from her wrecked body to be transplanted onto Mary Frances Beaudine’ healthy body, which had been left with a wrecked mind. When that person wakes up thinking she is Julia North, a dilemma arises. Either Mary Frances survived with drastic psychological changes, or Julia survived with a new body. The former would hold with bodily identity; the latter, with psychological continuity or memory theory. Though the world concedes that it is Julia who eventually wakes up, neither Mary Frances’ husband nor Weirlob does. Weirlob thinks Mary Frances has simply woken up deluded, thinking she was someone else due to having memories that are not hers. Both of Perry [1977]’s other characters, Miller and Cohen, disagree. Memory theory, they defend, explains how it is possible to judge as to one’s own identity, without having to examine one’s body. Also, it explains the importance of personal identity. But, to think the memories of the new person were the ones from Julia, we need to cling to the fact that it is her brain. If not, they could simply be fake, deluded, the result of hypnosis. But that brings us back to bodily identity. Others have argued, alongside Perry [1977], for the necessity and sufficiency of bodily continuity for personal identity, such as Penelhum [1971], who holds that it provides a public standard for determining whether some thought or feeling or action belongs to the biography of a particular person. There is a similar conflict of criteria as in the Ship of Theseus (§2.8) in personal identity, often illustrated by the brain duplication scenarios. Parfit [1984] or Wiggins [1967] are good examples of this. They suggest that it is logically possible for one person to split into two bodies, 3.1. Brain duplication against personal identity 65 each with an equal claim to be the original person. As we saw in §2.8, some theorists will claim that if both claimants or competitors, have equal claim to be the original (the Ariadne), then neither is. On his side, Parfit [1984] argues that if one competitor is inferior, then the other will win and be counted as the original. But Wiggins [1967] is convinced that the remodeled ship is the better option —it exhibits, he maintains, a greater degree of spatio-temporal continuity with the original. However, it is unclear why the preservation of identity through spatiotemporal continuity (the Aegean Ship) should take precedence over the process of disassembly and reassembly (the Piraean Ship). In Naming and Necessity (Kripke [1980]), Saul Aaron Kripke argued that an object made out of a particular material composition could not have been made out of a totally different material composition. So, on the matter of the Ship of Theseus, if there could be a world in which the ship was built of different planks (the Aegean Ship), that the Ariadne was not dependant of, then if another ship was made of the original planks (the Piraean Ship) the Aegean Ship would not have the claim to be the Ariadne. This kind of reasoning will be common throughout this chapter —a property which could be different under alternative circumstances or in different possible worlds is a contingent property (§2.4). If we grant that the Aegean Ship and the Ariadne cannot be the same ship, we seem to have the solution to the paradox. By the Kripkean argument, only the Piraean Ship has any claim to being the original ship. But Deutsch and Garbacz [2023] points out that the success is short-lived, as it leaves us with an additional paradox: suppose that the Aegean Ship eventuates from the Ariadne by replacing one part of the Ariadne one day at a time. There seems to be a widespread agreement that replacing just one part of a thing with a new, exactly similar part preserves the identity of the thing. If so, then, by the transitivity of identity, the Aegean Ship and the Ariadne must be the same ship. It follows that either Kripke [1980]’s argument is incorrect, or replacement of every single part (or small portion) does not preserve identity —which brings us back to mereological essentialism, which we saw in §2.3. Kripke [1980] flees the scene by denying that his arguments appliy to the case of identity over time, deeming change irrelevant in this case. However, Deutsch and Garbacz [2023] add, one may suspect that, if applied to it, Kripke [1980]’s reasoning will yield an argument for mereological essentialism —which brings us back 66 Chapter 3. Personal Identity to §2.3. Let us review the problem once again. Two ships are made at the same time: one sails, and the other stays to serve as the piece storage for when the first one has to be repaired. Which one is the ship? Neither? Not according to Kripke [1980]. It almost seems as though the process of “reparing” the Ariadne is really just an elaborate plot to dismantle and reassemble it —let us call the reassembly Arianna. And the Arianna and the Aegean Ship are the same ship, then since the Ariadne and the Arianna are distinct, the Ariadne and the Aegean Ship cannot be the same ship. However, this view is not free of criticism: Salmon [2005], for instance, argues in favour of the material of an object being able to be totally replaced without affecting its identity, while others, among which stands Wallace [2011a], fight mereological essentialism. Deutsch and Garbacz [2023] emphasise that the Weak version of Relative Identity (which we saw in §2.2) draws attention to the idea that two things may be equal relative to a third. The basic idea is that identity through change is not a matter of identity through successive, accumulated changes —this notion is at odds with both intuition and the Kripkean argument: through successive changes, objects can evolve into other objects. The three-place relation of identity does not satisfy Leibniz’s Law and is consistent with the view of a relativist like Gupta [1980]. Some have seen Locke [1689]’s prince and cobbler example as supporting the idea of relative identity. Shoemaker [2008] finds the notion indefensible and would prefer not to foist the notion onto Locke. He posits that Locke’s view needs coincident entities and that neo-Lockean views need them as well. But Shoemaker [2008] sees men unfit for being entities distinct from but coincident with persons. There is non-identity as well as identity, and to avoid relative identity, there must be some entity coincident with the pre-transfer prince that is nonidentical with some entity coincident with the post-transfer cobbler-body person. Neo-Lockeans see no issue here: each person is coincident with a human body, and the person and their body are non-identical because they have different persistence conditions (Shoemaker [2008]). Cole [1991] sides with them for similar reasons, as we saw with Searle and Polo, or the case of physical and virtual machines (§2.2). A person is an attribute of a body. (David Cole in Artificial Intelligence and Personal Identity) 3.2. Psychologists are also getting involved 73 scandalous (and known) theories —i.e., psychoanalysis— hint at this ambivalence. However, notorious as he was, Freud is rarely cited in modern works on identity. This may be due to his focus and aspiration that psychology be considered a natural science, based on an organic substructure (the brain), rather than on identity per se. Or maybe it is his perceived discredit among the authors who succeeded him. Still, we will begin this section by giving Gundersen [2023]’s defence of Freud’s perspectives on consciousness, the mind-brain problem, and the nature of psychological explanation a chance. The mind-brain problem is one of philosophy’s most debated problems. Though he was explicit we do not have a metaphysical solution to it, Freud seemed to believe that the mind and the brain were identical for most of his life and saw the existence of consciousness as being so apparent there was no need to characterise it (Gundersen [2023]). This is a position we will quickly find to be contested. His original topographical model of the mind distinguishes between conscious, preconscious (unconscious until given attention), and subconscious states. He poses that conscious thinking is the same as inner speech —thus, preconscious and unconscious states become conscious when verbalized in this form. This is a fundamental aspect of psychoanalytic therapy. Freud never doubted the brain is the substance of the mind. Moreover, according to Gundersen [2023], he implicitly subscribed to a kind of materialism or identity theory which claimed mental states to be identical to brain states —mind and brain as a unity to be described and investigated from both a psychological and neuroscientific perspective. Consciousness is the subjective side of one part of the physical (brain) processes in the nervous system, namely of the perceptual processes —though, in a Kantian2fit, he considered beyond human understanding why some physical events must be accompanied by conscious experiences. But even though consciousness cannot be explained by reference to brain processes, Gundersen [2023] argues it does not follow that the mind is not identical to the brain. Identity is an ontological relation and does not imply explainability, just as non-explainability does not imply non-identity. In this, Freud again shows his affinity to Kantian epistemology: perception does not necessarily reach the inner nature of physical objects. Gundersen [2023] observes that he seemed to back 2In short, the Kantian view on natural knowledge states that there is an essential limit to our knowledge of physical reality. 74 Chapter 3. Personal Identity the dual-aspect theory, which differs from the Cartesian dualism in that it does not presuppose interacting substances but contends that consciousness is an intrinsic and inseparable aspect of some physical processes. Famously, Descartes was radically suspicious of the veracity of all objects of experience. Only the self was exempt from this problem, as he thought it was not an object of any experience. Some computer scientists, such as Paul Smolensky, have viewed consciousness as a virtual machine (Cole [1991]), stored and emulated in hardware. Although not a dualist, he does not identify the self with some substance other than one’s body, Cole [1991] would agree with Descartes that one might exist while one’s (present) body did not. Rubin [2003], on the other hand, thinks we would be justified in calling a machine conscious the moment it is capable of responding to stimuli in a way that is indistinguishable from human responses; while Serrano [2024] argues that Artificial Intelligence will never form a consciousness or human intentionality, rendering it a mere algorithmic tool, dependant of its creators. Hume, with his understanding of the soul, shows a more nuanced problematic of apperception3than a simple dichotomy of inner and outer —the self is neither of external nor internal experience, as it is not an object of experience at all (Rosenberg [1981]). In the same vein, Robert Nozick’s Anarchy, State, and Utopia (Nozick [1974]) offers a thought experiment to disprove the claim that the pursuit of pleasure and the avoidance of pain are the only factors that contribute to a person’s well-being. Nozick wrote extensively on many subjects, and although his cited work in this thesis focuses mainly on political thought, —his contribution serves as a kind of warm-up for §4, where politics become more prevalent. Now, he presents us with an experience machine that feeds sensations and impressions into people’s brains. Its users may choose what they are to be induced with, and when they are plugged into it, they may experience anything they want, blissfully unaware that the experiences are not genuine. He then asks whether people would be inclined to accept to spend their lifetime floating away, tied to an array of electrodes. Nozick [1974] asserts that most would say no, and he lists three reasons for it. The first is that we want to do certain things and not just have the experience of doing them. The second one is that we want to be a certain way, to be a certain person. He writes: 3In psychology, becoming conscious through the subsequent reflection of an already experienced perception. 3.2. Psychologists are also getting involved 75 Someone floating in a tank is an indeterminate blob. There is no answer to the question of what a person is like who has been long in the tank. Is he courageous, kind, intelligent, witty, loving? It’s not merely that it’s difficult to tell; there’s no way he is. (Robert Nozick in Anarchy, State, and Utopia) Thirdly, plugging into an experience machine limits us to a manufactured reality. And yet, Nozick [1974] points out, the immense pleasure the person would feel for the rest of their life in the experience machine would be far greater than the amount of uneasiness that they would feel before being submerged. Nozick is showing that many people have the desire that their experiences actually be real. And this is, in fact, a desire that is completely independent of the notions of pleasure and pain. He explores this in The Examined Life (Nozick [1989]), where he argues: What we want and value is an actual connection with reality. [...] To focus on external reality, with our beliefs, evaluations, and emotions, is valuable in itself not just as a means to more pleasure or happiness. (Robert Nozick in The Examined Life) But let us go one step further. We have begun to move towards our experience of reality and of ourselves in it, which Nozick [1989] says we must inhabit in order to be someone. In that vein, we find Arendt’s representation of the world as inner dialogue (T¨ommel and d’Entreves [2024]) and her assertion that only those who actualize the difference between the world and its inner representation in their identity allow conscience to emerge as a by-product of consciousness. We will now focus on apperception, which is mental perception or introspective self-consciousness. This is not to say that self-consciousness is a widely accepted notion. As we have seen time and time again, the landscape of discussion is so splintered it almost seems fractal in nature. Behaviourists, who focus on observable behaviour as the primary object of psychological study, tend to deny the phenomenon altogether (Rosen [1974]). Their justification is that in the preanalytic state, we suffer from the illusion that we are self-conscious or have an inner experience. 76 Chapter 3. Personal Identity They see the sources of behaviour as external (environmental) rather than internal (mental), and urge that mental concepts be translated into behavioural ones, or eliminated if no translation is available. But, Rosen [1974] criticises, it is not possible to make a complete or even minimally satisfactory translation from an internal language to an outer language, so the most it accomplishes is to make it illusory. Plato accepts the phenomenon of self-consciousness yet denies the possibility of selfknowledge (Rosen [1974]) —he found conundrums in his way he deemed insurmountable. To Hegel, self-consciousness is a process of development which culminates in self-knowledge, an interpretation benefited from post-Platonic thought (e.g., Aristotle, Spinoza, Kant). Still, at least some of the Platonic conundrums continue to be present in Johann Gottlieb Fichte4—both hold there is no knowledge of the self, but in Fichte’s self-consciousness is a first principle from which other determinations can be deduced. At the same time, in Plato, it is something that may be inferred in part from the fact that intellectual intuition grasps forms, not the self (Rosen [1974]). Plato does not define the mind in a technical sense because, to him, it has no form of its own. Aristotle comes closer to a sense of form as activity, but his noetics5strain to keep distinct the form and the mind which has assumed said form in thinking it. In Plato, the mind, even though open or receptive to, is nevertheless distinct from the form it sees or knows (Rosen [1974]) —the soul (or mind) cannot know itself because in the decisive aspect (the capacity of knowledge) it is not knowable. Much closer to our time, and as a reminder of the intertwining of thought on personal and cultural identities, we find Ernst Cassirer6further develops this idea, emphasizing the symbolic nature of human cognition and how it shapes our understanding of reality. Heidegger [1927, 2012]7shifts the focus to the existential structure of human understanding, arguing that our being-in-the-world precedes and shapes our cognitive activities. Therefore, one’s identity both expresses one’s traits and is fundamentally shaped by one’s interactions with the world and others. 4Fichte (1762-1814) was a German philosopher and a key figure in the development of German idealism, known for, among other things, emphasising the role of self-consciousness and subjectivity in understanding reality. He was Hegel’s (1770-1831) senior by eight years. 5Noetic means of, relating to, or based on the intellect. 6Ernst Cassirer (1874-1945) was a German philosopher who developed the philosophy of symbolic forms, emphasising the role of symbolic systems in human cognition and culture. 7Martin Heidegger (1889-1976) was a German philosopher known for his existential and phenomenological explorations of the question of being, most famously in his book Being and Time (Heidegger [1927, 2012]). 3.2. Psychologists are also getting involved 77 On Hegel’s side, the initial discovery of self-consciousness in the struggle for recognition is continuously being lost, acquired, and lost again until the individual achieves self-knowledge. But if self-consciousness is the result of identity-within-difference of the subject and object, it would seem that it must also precede that result. The subject must first see itself in order to give an account of itself as an object, which is the process of identification between subject and object (Rosen [1974]). If it sees itself as an object, as Plato leads to, then it is initially separate from itself as a subject, and no explanation can be given to the act of self-consciousness; if one follows Fichte and the subject sees itself as a subject, then it presupposes itself, and again, no account is given of the origin of self-consciousness. Therefore, Rosen [1974] notes, the act must have been completed before it takes place; otherwise, there would be no subject of the appropriate kind to be discovered. In Hegel, self-consciousness is the self both opposed to and comprehending itself, and its structure is the identity of the individual and the Absolute. Jean-Paul Sartre defends that all consciousness is consciousness of something: it transcends itself to reach an object and exhausts itself in the process (Rosenberg [1981]). This is a theory of non-self-identity, which entails that the self is not identical to its own appearance (Jordan [2017]). According to Rosenberg [1981], any object of Sartre’s consciousness is necessarily posited as something other than the consciousness that is directed towards it. It can no more take itself as its object than a mirror can take itself —a consciousness that is aimed at self-consciousness as an object of reflection would be analogous to a map wanting to represent itself as a map. But a map that could represent itself as a map would have to be both a map and its own territory (Rosenberg [1981]). Sartre criticizes the idea, noting that it would lead to an infinite regress. The concept of infinite regress gets thrown around a lot. Taking reflection8in the comprehensive sense, Rosen [1974] asserts, it seems clear that its presence is enough to debar self-knowledge. Phenomenologists, linguistic analysts, and materialists affirm that one cannot be immediately conscious of oneself because the effort to grasp the self leads to an infinite regression of previous states of the self. If, however, a person remembers what they were once like, they must recognise themselves in these memories and hence must be acquainted with the self they are in the process of identifying —thus, Rosen [1974] concludes, retrospection is a special case of 8A central concept in Fichte. Reflection is presented by Kant as a condition for the formation of concepts in general (de Figueiredo [2020]). 78 Chapter 3. Personal Identity reflection. But the reflection problem has only begun to form before our eyes: discourse, even as it identifies itself, is sufficiently different from the form with which it identifies to contradict the utterance. Rosen [1974] claims that we require the element or dimension of self-consciousness in addition to the form referred to as defining the self in order to account for the reference of identification. But this means that in order to account for self-consciousness, self-consciousness must be presupposed. Arguments like this demonstrate the impossibility of conceiving the self as a formal structure. The belief that the mode of being of a self is an activity —cogitatione, according to Descartes— is one that Sartre shares with the scholastic-Cartesian tradition (Rosenberg [1981]). As argued thus far, the existence of consciousness must precede its essence, since consciousness cannot exist prior to being. Still, its being is the source and condition of all possibility. Thus, Sartre concludes that any positional (i.e., transcending itself) consciousness of self must necessarily rest on a foundation of non-positional (pre-reflective) consciousness. It is a selfconsciousness without being a consciousness of the self. T his is what he calls nihilation: the perpetual activity of constituting the object of noetic consciousness as other than the constituting consciousness. To constitute a consciousness as other than its objects is a condition of the potential for such noetic activity, that is, of the potential for a consciousness of itself as other than its objects. But Rosenberg [1981] challenges this concept, finding Sartre’s argument flawed in its assumption that a consciousness lacking self-awareness would inherently be ignorant. He argues there is no evident absurdity in the notion of a positional consciousness of an object that is not simultaneously consciousness of being conscious of that object. Jordan [2017] offers Sartre a hand in this regard without overtly rescuing him. She argues that an implicit argument by Sartre claims that consciousness is non-identical —i.e., it lacks selfidentity. Sartre does not simply intend the diachronic claim that I, the self-conscious subject, am not identical with my past or future consciousness —but also the synchronic claim that consciousness fails to satisfy the principle of identity: ∀x(x=x). But his explicit argument for it seems weak, as we have begun to see with Rosenberg [1981] and Jordan [2017] reaffirms and does not explain how pre-reflective self-consciousness relates to the experience it apprehends. Iterative pre-reflection sees the subject aware of its experience and aware that it is aware; 3.2. Psychologists are also getting involved 79 non-iterative pre-reflection finds the subject aware of its experience but unaware that it is aware. Iterative theorists defend there is an identity between pre-reflection and pre-reflection’s appearance itself and to itself —they are therefore committed to the strict self-identity of the subject. Since Sartre holds that pre-reflection is non-self-identical, according to Jordan [2017], Sartre is a non-iterative theorist. Sartre’s argument for non-self-identity has three premises. The first one is that subjectivity is just pre-reflective self-consciousness, and conversely. The second is the most controversial one: there cannot be a strict identity linking consciousness with its appearance itself and to itself. Iterative theorists are committed to denying it (Jordan [2017]). Finally, the third premise states that a subject must somehow appear to itself —granting the first premise, one can say a subject must, in some sense, be its self-appearance. A non-self-identical subject, though it will not be identical with its self-appearance, it must in some way be its self-appearance. But Jordan [2017] observes that the explicit argument is too concise to convince. It seems not to inform us about the nature of this non-self-identical subject other than saying it is non-self-identical: it only tells us that consciousness cannot be identical with its self-appearance. The implicit argument found by Jordan [2017] sees the second premise changed to selfconsciousness is non-iterative, a position the original one already alluded to. According to it, the consciousness of an experience cannot be conscious of itself as a consciousness of experience. This has two important consequences. Firstly, the non-iterative consciousness of the experience cannot be identical with the experience. Then, it follows that the non-iterative consciousness of the experience cannot be self-identical. If we assume that self-consciousness is non-self-identical, then we must say that I am my consciousness of the experience in a sense that acknowledges that the subject is somehow dual, in some sense ruptured from itself. In this, we have found Hegel again. Jordan [2017] writes that the consciousness of the experience cannot, as a matter of structure, distinguish itself from the experience —the non-iterative subject cannot apprehend the experience as distinct from itself. So, the non-iterative subject must be the experience. That is, be what I am not. Sartre is but one of the many who have tried to pinpoint the nature of consciousness against Freud’s best wishes. This brief outline of his theory (and implicit theory) is intended to preface 80 Chapter 3. Personal Identity the utter complexity of the discourse on self-consciousness, which is just one of the crutches philosophers have leaned on to hunt down personal identity. As this section progresses, we will be looking at a few more representatives on the battlefield, starting with a trip back in time to §2.6. To free ourselves from self-interest and fear of death, Penelhum [1971] notes Parfit’s suggestion that we think of our lives not as one unit but as divided into the lives of successive selves, the passage from one to another marked by significant differences in lifestyle or character, or by loss of memory: a lessening of what he calls psychological connectedness. Merricks [1999] opposes Locke, posing that if persons endure (as he defends), then it cannot be that personal identity over time should be analysed in terms of psychological continuity. He merits Locke for his almost complete hegemony of it in contemporary analyses, most of which understand personal identity over time in terms of psychological continuity because of the conviction that to be a person fundamentally involves having some array of psychological traits. Psychological continuity is the ancestor of psychological connectedness, which entails a person at one time sharing psychological states (e.g., memories, beliefs, desires, and intentions) with a person at another time in overlapping chains. This aligns with Parfit’s concept of the preservation of personality, temperament, beliefs, values, and goals through discrete connectedness (Penelhum [1971]). It also involves a person at one time having memories, apparent memories, or quasi-memories (Q-memories) of a person’s experiences at an earlier time. Incidentally, these concepts could also apply to BDI-based Autonomous Agents, which, as established in §1.2, are designed to replicate human reasoning and decision-making processes. Their representations of knowledge about their environment (beliefs), goals (desires), and commitments to courses of action (intentions) closely parallel the psychological states involved in human psychological continuity. Thus, it is conceivable that a form of psychological continuity could be modelled in them, potentially allowing for a computational representation of persistent identity over time. Let us recall Wallace’s worm-like objects (§2.3, Wallace [2011b]), Yagisawa’s stage-ridden concrete entities (§2.5, Yagisawa [2010]) and Lewis’ Counterpart Theory (§2.5, Lewis [1991]). All of them involve Four Dimensionalism (perdurantism, the view that a persisting object at any one moment is not the object as a whole), which finds person stages in people as if they were temporal parts. Those person stages are considered related as if they were links in a chain, so 3.2. Psychologists are also getting involved 81 personal identity may be spread over time and hold its unity and distinctness. According to Four Dimensionalism, a person’s existence at a time is just their having a proper part, a person stage, that exists wholly present (not spread out in time and without temporal parts) at that time. In endurance ontology, Merricks [1999] writes, psychological continuity would mean that a person existing now is related to a person at a future time by saying they have certain psychological states now, and those states are related to the states the person existing at the future time will have. So, it is not that the person is related to themselves, but rather the way a person is at one time or another. An endurantist like Merricks [1999], unlike the four-dimensionalist, holds that a person’s existing at a time just is them (the person themselves) existing wholly present at that time. So, both doctrines agree that personal identity over time is analysed as a relation between something that exists wholly present at one time and something that exists wholly present at another. However, the endurantist thinks that the relation in this analysis is numerical identity and that it relates to a person and themselves —while the four-dimensionalist denies it. But we have stated that the endurantist would accept an account of psychological identity between non-identical psychological states... so it cannot be that personal identity over time just is said relation. The only hope for reconciliation, Merricks [1999] thinks, will be in a sort of psychological continuity that holds between a person and themselves. So the endurantist makes a minor correction to the relation they defend, and suddenly, a person existing now is (unbranchingly) psychologically continuous with a person existing at a later time only if a person’s current psychological states are related to those that a person existing at said time has at that time. But this, even when holding between a person at one time and that same person existing at another, is something other than numerical identity. If persons endure, no attack on the value of any unbranching relation should cast any doubt on the value of personal identity over time —and psychological continuity of every sort is a relation, no matter what the details of the suggested continuity are. This argument thereby requires, Merricks [1999] argues, that personal identity over time is distinct from both numerical identity and identity over time. Merricks [1999] gingerly disrupts nearly everything we have commented on so far and will go on to overview. He rejects in one fell sweep Cartesian substance-based continuity: it implies, he 82 Chapter 3. Personal Identity asserts, that psychological continuity is necessary and sufficient for personal identity over time, but this does not imply that personal identity is reduced to psychological identity. He then focuses on the constitution (§2.4), which could serve to explain the relation between personal identity and psychological continuity (we turn, again, to Clay and Goliath) as amounting to something other than necessary and sufficient conditions. But this something other is much too undetermined for Merricks [1999]’s comfort. Mere necessary and sufficient conditions for personal identity do not, he concludes, amount to a genuine reduction. To close the argument, Merricks [1999] takes a stab at Parfit’s attack on the claim that identity matters in survival and renders it unsuccessful. A s Parfit himself recognises in Parfit [1984], the view that personal identity does not matter in survival is counterintuitive and very difficult to believe. Weirlob, Miller and Cohen already explored for us in Perry [1977]: if I were to be destroyed and a duplicate of me took my place, should I be putting all my affairs in order, in preparation for an untimely death? If identity does not matter in survival, then the answer to this question does not matter. Parfit [1984] says that, even if the duplicate were not identical to me, it would still have what matters —psychological continuity— and I should anticipate its experiences just as I would anticipate my own. As we saw in §3.1, Parfit [1984] defends that if one were to be subjected to a case of fission, both products would be psychologically continuous to that person. If one were to avoid fission, then one would be psychologically continuous to one person in the future. But would this show that it is psychological continuity and not personal identity that matters? Merricks [1999] answers that only if personal identity is just unbranching psychological continuity. However, he has shown that an analysis or reduction of personal identity over time in terms of unbranching psychological continuity is simply not possible given the thesis that persons last over time by enduring. If persons endure, then Parfit’s argument is false. Francescotti [2010] follows a very similar line to Merricks [1999], but focuses on a different difficulty. He shows that the psychological continuity approach —as both endurantists and perdurantists frame it— conflicts with the doctrine of the Necessity of Identity9, which we saw in §2.2. Francescotti [2010] notes that memory relations are not enough to link two bodies across 9The idea that true identity is never contingent (see §2.2). 3.3. What memory says about us 89 to memory but a constituent of it as well. Then, memory would be a complex mental state consisting of a conception of a past event in which one was either a witness or agent and a belief that said event happened. This interpretation is better than the last, Coppenhaver [2006] grants, as it explains why memory represents events as having the special quality of being in the past —the former explains why we believe that events we remember are in the past, but cannot explain why memory represents these events as past because it does not regard the belief that event occurred as constitutive of memory. But has this last shot backfired? Does this notion of preservation not appeal to the storehouse metaphor, which even Locke attempted to bypass with his notion of mind? Moreover, like Locke, Reid holds that memory is partly constituted by belief... yet Reid manages to dodge his own bullet by insisting that memory is not a current apprehension but rather a preserved past one: it is not directed towards any present or past perceptions or events, stored or otherwise. Rather, Coppenhaver [2006] explains, memory is directed towards the events presented in previous perceptions —because perceptions (apprehensions) are never the objects of memory; they need not be stored for use as an object of memory. Likewise, the belief that is an ingredient in memory is not a belief about any present or past perceptions or apprehensions. If it were, Reid’s theory would be just as circular as Locke’s, Coppenhaver [2006] asserts. The force of this theory is best illuminated by contrast to the theory of ideas, which he so ruthlessly riddled with bullets. But let us now pan out of this stand-off and take an overview of the whole field of memory theory. Miri [1973] explores Shoemaker’s insistence that without the special access afforded by memory to one’s past, no purposeful behaviour (which is central to the notion of a person) would be possible. Moreover, for him, the obligatoriness of a person witnessing an event to remember it and the causal nature of memory leads to the natural conclusion that memory is constitutive of personal identity. But if that means that it is a necessary condition of personal identity, Miri [1973] ponders, then amnesia seems to increase its level of danger from life-altering to self-extinction: would one, faced with the knowledge that they were about to lose all their memories, think that they would cease to exist? It would be horrifying, she admits, but some semblance of hope to rebuild their life would remain. Furthermore, the issue of memories being 90 Chapter 3. Personal Identity true or false, which we saw with the modern problem of circularity (3), gives wake to the notion of a criterion prior to memory, which would judge whether a potential memory could happen or not. One could argue that the existence of a causal chain beginning at the time of the experience could be enough to know that someone is potentially capable of remembering something. The issue is, Miri [1973] notes, the possibility of memory implantation obliterates the necessity of a person having access to their, and only their, past —thus turning this criterion contingent (3.2). In a study on the risk of using advanced AI-based systems in police interviews by Chan et al. [2024], it was found that interactions with a generative chatbot significantly increased false memory formation, as the subjects were misled through the interaction, though a less significant (but still effective) result can be achieved through pre-scripted chatbots or even just a survey mined with deceiving questions. Even more troubling, Chan et al. [2024] found that they not only persist but also maintain their perceived credibility over time. This concern is exacerbated by AI-based models’ known yet unresolved tendency to hallucinate or generate false information, either intentionally or unintentionally. There is a troubling danger in sycophantic AI (its tendency to provide responses that align with user beliefs rather than the objective truth), which creates an echo chamber effect where users’ existing biases or misconceptions are validated and reinforced. It thus becomes a real-life example of what we have treated as mere hypothetical. This problem could be solved with the exstinctionist dream of exchanging biological life for copper and silicon, which would dramatically expand memory and change its character to non-deciduous (Rubin [2003]) —but this would be at the cost of a host of more complicated questions about the continuity of the self from one’s body to a software receptacle. Wiggins [1967] contends that to be a person is to be capable of believing and disbelieving things on evidence, which in turn requires the possibility of memory of experiences. Shoemaker [1966] adds that a creature that lacked the ability to have this knowledge of its own past history would be incapable of knowledge of any kind. But could a person’s memory of such experiences be immune to misidentification, as it would need to be if it were to lead to knowledge? First of all, Miri [1973] argues that nothing is inconsistent in supposing that a being can remember an experience without necessarily remembering themselves experiencing it, as the experience might have been someone else’s. Yet, they could have a standard procedure for finding out whether the 3.3. What memory says about us 91 experience was theirs or someone else’s. If Shoemaker’s memory theory is correct, to remember an event is at least to locate it within a certain unique spatio-temporal area: the one occupied by one’s own history —assuming one has a body and that body shares that history. But the claim that in memory one necessarily has an infallible access to one’s own identity cannot be correct, Miri [1973] concludes, as there could be worlds in which to remember would be to specify a certain more or less limited area traced by the causal trace —also, there would be very good empirical grounds for believing that within this limited area, no reduplication of things and places occurs, which would make identification of anything in the past impossible. Still, there seems to be absolutely no reason why purposive behaviour should be impossible. Selfconsciousness, knowledge of any kind, and intentional action depend on memory. And it is at least prima facie evidence of personal identity (Miri [1973]). In §3.2, we encountered the battle between psychological and bodily continuity. Now, we find one between memory and bodily continuity. For that, we go back to Shoemaker’s Brownson: Brown’s brain in Robinson’s body. Where we left them, Brownson had retained Brown’s past experiences, as well as his personality traits. But what if he instead kept Robinson’s mannerisms, interests, likes and dislikes? If persons are where memories are, Miri [1973] notes, Brownson should still be identified as Brown —a memory theorist would presumably die on this hill even if Brownson were composed of half each man’s brain. But he may only feel vaguely familiar with the events now in his head, and with the people that Brown knew, with no strong belief, those memories are his —if he retains Robinson’s ways, Miri [1973] thinks this to be plausible. For the sake of argument, Miri [1973] separates memory from character formation, allowing Robinson to identify as Brown while having undergone a personality change. But let us suppose some of Robinson’s memories creep in. Wouldn’t it be arbitrary to still go with Brown, who still has most of his memories in Brownson? Memories are supposed to be uniquely and exclusively of the individual, as we have seen. Miri [1973] challenges: would it not be more logical to go with Robinson, with whom Brownson shares body? And, what if Brownson suffered from complete amnesia? Even if he gradually gained personality traits similar to Brown’s, bodily continuity theorists still would side with Robinson. One could argue that Brownson can be identified as Brown because Robinson has in him everything causally 92 Chapter 3. Personal Identity necessary for the eventual regaining of Brown’s memories. But Miri [1973] asseverates that a memory theorist could not hold that because the physical conditions of memory are yet to be specified and, for them, it is memory and not just brain identity, that is the criterion of personal identity. This is an argument found in Perry [1977] as well. Even if he were to regain some of Brown’s convictions and therefore believe he was Brown, Miri [1973] makes the case that Brownson could be fed fake memories of another individual and be convinced he was that individual in the same way. One might say that this could be solved by verifying Brown’s memories against reality, but the degree of his assurance may be the same with both real and fictitious memories. This might tempt us to bypass belief entirely, but Miri [1973] warns that we would be misplacing personal identity, since if Brownson’s memories need not be accompanied by the belief that they are memories, and that they are memories of experiences of his past life, doubt would remain. Anything could be fed to Brownson, as long as it was based on a factual occurrence. Moreover, if memory theory had to allow the converse (both allowing and giving equal weight to more than one non-identical memory claimant), then it would violate the transitive, symmetric and reflexive relation of identity (§1). Shoemaker, for instance, following Williams and Nozick (Nozick [1974], Nozick [1989]), discusses the possibility of a device which records the state of one brain and imposes it on a second brain by restructuring it so that it has exactly the state the first one had at the beginning of the procedure, thereby transferring both memory and character traits (Noonan [1998]). Thus, the genuineness of memory could not guarantee personal identity, nor could subjective belief. This, Miri [1973] points out, is a very serious shortcoming of memory theory, which ultimately renders memory unsuitable as a criterion of personal identity —though Wiggins [1967] believes this could be overcome by a due recognition of the fact that memory is causal notion. Miri [1973] concludes that our judgement of personal identity would not be impaired if our criterion of identity was bodily continuity. 3.4. Our place among the animals 93 3.4 Our place among the animals There are two major strands to the history of personal identity: Locke’s faction, building on psychological continuity, and rejections of Locke and neo-Lockean views, which gather soul theory (e.g., Butler [1736]), bodily identity (e.g., Perry [1977]) and animalism. Shoemaker [2008] argues that these reactions exist only as opposition to Locke and that the ideas would have been considered old even back when they were presented. Animalism, the dominant position in the anti-Lockean movement, holds that persons are human animals and that the persistence conditions of persons are biological rather than psychological (Shoemaker [2008]). Thus, it argues that persons are entities distinct from animals, with which they are merely coincident —without denying the possibility or even the actuality of non-animal persons. It defends that we are essentially animals, Noonan [1998] adds, which implies that for each of us, there is a certain animal, the condition for whose survival is the condition for our survival. This position opposes that of Locke and his tripartite ontology (§3). With his notion of thinking substance nowadays widely rejected, neo-Lockeans only observe the theoretical possibilities of a human persisting while different persons may inhabit it at different times and the inverse case, which sees a person persist (Noonan [1998]). As per Shoemaker [2008], animalism is not a special case of the issue between materialist and dualist accounts of the mind, which we saw in §3 —neo-Lockeans, though most of them materialists, admit that the mental states of affairs they defend constitute personal identity are physically realised. We have followed a notorious neo-Lockean in this chapter: I have long been, and remain, a neo-Lockean. (Sydney Shoemaker in Persons, animals, and identity) At its core, the discussion revolves around the simple observation that where there is a person, there is surely a human animal or human organism. The separation of terms (person and human animal) is neo-Lockean, thus superimposing them —animalists are more decisive and claim that the person is the human animal. This, in turn, prevents body-switching scenarios. Shoemaker [2008] posits that Locke [1689] was thinking of men as having the persistence conditions of 94 Chapter 3. Personal Identity animals in his denial that the cobbler-bodied person was the same man as the prince. Rejecting that persons are identical to human animals is perhaps not as counterintuitive, Shoemaker [2008] defends, as denying that they (at least the males among them) are men. But it is still, he admits, prima facie counterintuitive. One last time, we go back to Brownson’s case in Miri [1973], where the animalist will claim that no human animal survives the procedure which takes Robinson’s body and Brown’s brain to create Brownson because the identity of an animal is not determined by the identity of its brain —the same animal can receive a new brain, just as it can receive a new heart, and, as Noonan [1998] argues, the preservation of an animal’s brain is no more tantamount to the preservation of the animal than the preservation of its heart. They will argue that, if neo-Lockeanism were to be correct, it would be logically possible that one should have a history distinct from that of any animal. That is, Noonan [1998] clarifies that not all persons that at some time coincide with animals would be necessarily animals, which would follow that one would not be an animal. Both Parfit and Van Inwagen [1990] argue that neo-Lockeans are not committed to such a possibility. Instead, they similarly claim that the identity of human animals is constituted by that of their brains, as the brain governs most of the body in several ways (Noonan [1998]). Moreover, Parfit holds that, in the case of a disembodied brain, sustained in a vat of nutrients and connected to a computer which feeds it, experiences indistinguishable from those of a normally embodied person10, the animalist might find difficulties. Either they deny that preserving the brain involves conserving a conscious being, or they accept that it does but deny that it is an animal. If the brain is transplanted into a new skull, Noonan [1998] follows Parfit’s challenge, the resultant entity will be an animal. But if there was no animal present before, either the conscious being that was there has disappeared, or it still exists and shares its thoughts with an animal. Moreover, what if Brown and Robinson were dogs? It could not be said that Brownson was the same person as Brown. As the same what, then, would one be tempted to identify Brownson and Brown? So, it seems plausible that neo-Lockeanism can also be animalist. In fact, Noonan [1998] observes that the Lockean account of personal identity (against Locke’s will) can be thought of 10Commonly known as the brain in a vat thought experiment. 3.4. Our place among the animals 95 as an explanation of the Lockean notion of identity of life, which, according to Locke, is what determines animal identity. Parfit goes on to argue that in cases where it is indeterminate whether a person survives (brain state transfers, complete cessation of all psychological continuity), it is also indeterminate whether personal identity diverges from animal identity. Alternatively, Shoemaker suggests that neo-Lockeans should accept persons are not animals, in the sense that they are not identical with animals, but insist that they are animals nonetheless in the sense that they share their matter with animals. But Noonan [1998] points out that this seems to imply that animals, even those that coincide with persons, are not thinking intelligent things, but merely share their matter with thinking intelligent things —yet animals, and in particular human animals, are our paradigm of what it is to think. As we began to see in §3.3, Shoemaker [2008] argues that the notion of personal identity closely relates to the characterisation of the functional roles of mental states, which generate successor states on the same subject, which will in turn do the same. This builds a chain of mental states and accompanying behaviours and includes the generation of memories, thus exhibiting all the sorts of psychological connectedness and continuity that have been thought to constitute the persistence of mental subjects. Of course, he concedes, it would seem circular to define psychological continuity in terms of functional roles and then define personal identity in terms of psychological continuity. So, he proposes to simultaneously define a number of states in terms of their relations to one another and of synchronic and diachronic unity, which can then be used to define personal identity. Then, assuming that mental states are realised in the cerebrum11, Shoemaker [2008] holds that were it to be transplanted to another body, the recipient would be the same person as the donor due to the psychological continuity found in the re-stationed cerebrum. This, we already saw in §3.3. But allowing this does not relieve the neo-Lockean from the obligation to accept that persons are coincident with non-persons: it remains true that they coexist with their bodies and with biological animals that are not identical with them (Shoemaker [2008]). 11The choice to refer to the cerebrum rather than the whole brain follows Olson’s point that the lower brain is the biological control centre of the human body and is compatible with his animalist view that the transplant of the whole brain would bring the person into the new body, not because it would provide psychological continuity, but because it would yield biological continuity carried by the brainstem. If only the cerebrum was to be transplanted, the person would stay behind as a human vegetable. 96 Chapter 3. Personal Identity Similarly, the too many minds (or, the thinking animal) problem arises in the case of nonhuman animals (Noonan [2010]). The denial that a brain transplant, or even a gradual replacement of organic substance without loss of consciousness, maintains the same rational conscious being applies to humans as well as dogs or whatever other animal. Faced with this problem, Cole [1991] does not think it possible to prove that there is actually a person (Polo,§2.2) who understands Chinese in Searle [1980]’s Chinese Room argument. In psychiatric terms, Cole [1991] admits that what he is proposing is very close to a case of multiple personalities, which he describes as having the essential feature of the existence of two or more distinct personalities taking control of the individual who hosts them at different times. His view is close to the neoLockean (more precisely, the functionalist) in claiming that persons or minds are more abstract than the dualist’s person with a body —a position implicitly held by Locke [1689]. Both animalists and neo-Lockean are vulnerable to it, Shoemaker [2008] observes. The latter, in holding that animals are persons and that they also share mental states with a person, leaves the body too crowded. Locke [1689] had also suggested that a single substance could be the home of more than one person if there is no psychological continuity —although this refers more to a kind of sequential parade of persons as if they were hermit crabs inhabiting a shell, rather than a coincident crowd. In defending that sharing the same physical properties entails sharing the same mental properties, the former gets the undesired result that the body shares the person’s mental properties, so it hosts too many minds. Noonan [2010] writes that, according to the psychological continuity theorist of personal identity, persons are not human beings. If human animals, in addition to persons, are thinkers, they must be persons as well since their thoughts have whatever complexity and sophistication any ordinary definition of a person would require. Noonan [2010] argues that only persons are objects of first-person reference and that all persons are psychological continuers. Thus, he defines a person as simply meaning “object of first-person reference”. To be a psychological continuer is to have a certain kind of history: of past, present and future. Another way to avoid the too many minds problem, put forward by Shoemaker [2008], is to deny that coincident entities must share all the same physical properties, and so they will be entities that have different persistence conditions. This brings us back to Gibbard [1975]’s example of Goliath and Clay and the paradoxical 3.4. Our place among the animals 97 observation that two entities which are identical in material constitution at all times may be distinct on the grounds of their being modal predicates true for one and not for the other. Leibniz’s Law (§2.1) appears to have come back to haunt us... until Noonan [1998] whips out Abelardian predicates, which are predicates whose reference can be affected by the subject term to which it is attached, and claims that the conflict with Leibniz’s Law is merely apparent. Lewis [1971] may be the most notable development (but not the only one; i.e., Gupta [1980] describes another) of the idea that modal predicates are Abelardian, thus becoming flexible enough to fit Leibniz’s Law without breaking it. This allows for Clay possibly being squished into a ball but Goliath not being able to, since each singular term differs in a sense, and hence there is no such thing as the property, which is the property of being a thing which might have been squeezed into a ball and not destroyed (Noonan [1998]). According to Lewis [1971], the property denoted by the predicate in the true sentence is has a counterpart under the lump counterpart relation which is squeezed into a ball and not destroyed, whereas the property denoted in the false sentence is has a counterpart under the statue counterpart which is squeezed into a ball and not destroyed. This could also be useful for Cole [1991]’s physical and virtual machines (§2.2). In a similar fashion, the neo-Lockean can answer the animalist by arguing that, given that there is a possible future in which a person is the donor of a set of transferred brain states, it follows that they are not essentially animals. However, it does not follow that they are not animals12, any more than we may conclude that Goliath and Clay are distinct (Noonan [1998]). According to Lewisian counterpart theory, is essentially an animal will denote the property has no counterpart under the personal counterpart relation which is not an animal when attached to the person but will denote the property has no counterpart under the animal counterpart relation which is not an animal when attached to this animal here —and these are different properties, Noonan [1998] adds, since the animal and personal counterpart relations incorporate different counterpart relations. This allows the neo-Lockean to consistently maintain both that personal identity is constituted by brain identity and that each of us speaks correctly, claiming to be identical with an animal. All they have to do is reject that each of us is essentially an animal, but this is entirely consistent with holding that all actual persons are animals. 12Note the difference between being essentially an animal and being an animal. The former denotes a fundamental and necessary aspect of one’s nature that cannot be separated from one’s identity, while the latter simply describes a current state or classification without implying it as a core, immutable characteristic of one’s being. 98 Chapter 3. Personal Identity Parfit holds that the notion of personal identity does not, or need not, have the importance philosophers have assumed it to have. In very broad strokes, we have found that the combination of physical properties with the experience of conscious states has traditionally been deemed enough to be considered a person. However, Penelhum [1971] points out that these conditions fail to distinguish persons from other sentient and deliberative organisms and omit mention of characteristics that are distinctive of persons —physical properties and conscious states are necessary but may very well determine the identity of animals. Wiggins [1967] challenges the Lockean conception of a person and its distinction against animals, which are regarded simply as living bodies. He writes: Person in my use is certainly a (non-biological) qualification of animal. (David Wiggins in Identity and Spatio-Temporal Continuity, Appendix, 5.4) Wiggins [1967] treats animal as a sortal hypernym that allows for a less individuated reference than person or horse does. The latter would amount to a sortal cross-classification —then, person may be a concept which only becomes fully determinate when one is told what sort of person, through utterances like man-person or even dolphin-person, etc. Interestingly, Olson [1998] tells us that E. J. Lowe maintains that dogs and dolphins are, or have, selves, even though they are not people —a very compelling example of the arbitrariness of most proposals on personal identity. This may be a response to what Jackson [2017] calls our innate Theory of Mind, which is the most rudimentary, foundational human psychology as studied in the cognitive sciences; our intuitive knowledge that other human beings have intentional states like our own. It is the psychological foundation of the human animal’s uniquely social identity. This Theory loses its balance in exploring imitation (§3) and, particularly, mirror neurons (i.e., motor neurons which only activate in mirroring an observed action) systems, which have been proposed as the biologically-based interindividual mechanism that enables it to operate. Harry Frankfurt argues that it is distinctive of persons to have what he calls second-order volitions, a sort of free will —to want one’s actions to be determined by one’s desires (Penelhum [1971]). Thus, freedom is something that an animal cannot have, and only a person can. In 105 what differentiates one from others. In many ways, Austin [2005] ponders, those characteristics are what from some perspective constitute culture. Nowadays, culture has developed in two significant directions: on the one hand, as an organic and intricately entity connected to social construction processes; on the other, as micro social groupings, such as street gangs or the touring rock band (Austin [2005]). As we saw in Austin [2005]’s quote in §1, there is a very notable issue in the field of identity: the confusion between identity and identification. This is not a phenomenon contained in social or cultural identity —the chaos we have trod through in §2 and §3 could be explained in a similar fashion, an observation to which Olson [1998] already alluded in §3.2 with his argument on the non-existence of the problem of the self. Philosophers have had the ambition to merge (or at least have not wanted to see the distinction between) the concepts of identity and identification, and this has given rise to the fractal factions of the debate, as if a mirror was broken into so many pieces it became little more than shiny dust. The plastic cards, PINs and formal identifications are but superficial markers of an assumed identity: features or characteristics that other people take to mean we are such-and-such, Austin [2005] points out. They are what could be called accidental (in an Aristotelian fashion, introduced in §2.5), or non-intrinsic, or whichever terminology countless authors have chosen before us. One thing is for certain, as we cited at the start of §3: identity only becomes a problem when it becomes uncertain. Austin [2005] places us in the era of postmodernism, which began around the second half of the 20th century and is still ongoing. It has created a world of multiplicities, diversity and change, demolishing the monolithic anchor points for understanding the self and others that ruled in the past. The period is characterised by the breakdown of the grand narratives, such as the continual improvement in the conditions of human life through progress. The disintegration of the notion of progress from the vortex of hope for the human species has created a void at the centre. Yin [2018] premises postmodernism on plurality and difference —on its attempt to take all privileges down and seek a more equal representation of gender, race, class, sexual orientation, culture, etc. But we ought to be wary of its grand banners, she fiercely argues: its celebration of difference does not necessarily lead to the empowerment of the marginalised. 106 Chapter 4. Cultural Identity Plurality is only meaningful when participants have equivalent representation, access to resources and opportunities, and a reasonable degree of equality in terms of power. However, it is evident that these conditions are not currently met. While plurality is celebrated, it is not acted upon in a manner that aligns with its principles, leaving it as a void gesture. Postmodernism emerged as a challenge of the dominant Western concept of the self as fixed, coherent, and autonomous. Instead, it emphasizes its fragmented, diverse, and pluralistic nature. In the postmodernist view, subjectivity is continuously shaped and reshaped through the dynamic interactions of discourse and language (Yin [2018]). However, while this discourse forgoes the rationality, order, and structure associated with modernity in favour of a more temporal and process-oriented understanding of the self, it remains deeply rooted in Western values. Paradoxically, Yin [2018] harshly critiques, it relies on the very methodologies of the tradition it seeks to dismantle, which limits its ability to establish a truly new epistemology. Austin [2005], deeply entrenched in Western thought, is also critical of postmodernist ethos. He situates the general malaise of this era in the sense of not belonging, of unreality, isolation and being fundamentally out of touch. The fading of the absolutist values of the Western Enlightenment, along with the collapse of foundational myths, traditional sources of authority and identity and political consensus, has led to it. According to him, in the postmodern worldview, there is no such thing as an essential me, no centring self-identity, and no in-born character. The individual assumes an identity and claims it for themselves based on a feeling or perception of commonality with others whose essential characteristics can be identified, named and compared, and ultimately given value. This process is largely unconscious, which is why Austin [2005] ventures that the incredulity that often greets a person’s coming out likely derives from the perceived audacity of someone consciously identifying and assuming a preferred identity. There is a focus on psychological consequences of identification, which speaks to the interconnectedness of cognition and emotion —Howard [2000] exemplifies it through prejudices, which may shape not only one’s own identifications but also one’s categorisations of others according to everyday commonsensical criteria, often linked to dichotomic relationships of belonging or not belonging. Davis et al. [2019] observe that identity and related emotional attachments can be a motivating force for social movement participation, driving individuals to contribute 107 to collectives. Cross-cultural research has found differences in the perception, expression, and desirability of different discrete emotions based on cultural differences in the self-concept and related cultural norms (Coleman and Williams [2013]). This suggests that cultural norms regarding self-construal often prescribe what an individual should feel, though discrepancies arise between the normative state and the actual felt emotion. In Hancock [2024], it is claimed that: [T]here is a psychological and social part which is key: it is us who give identity to this combination of sensations. [...] Our identity is built from many examples which may seem weak if isolated and scrutinized, but together show a frame which will generally not give us problems. (Jaime Rubio Hancock in Mocedades, el Partido Republicano y el barco de Teseo; my translation, from original Spanish) Austin [2005] sees social expectations as a disciplining mechanism attempting to force clear and clean identity choices. It also works to constrain the emergence of new identities —it is within the uncertain space of blurred that new identity possibilities are formulated. New forms and uses of language need to be drawn up to name them, and so language itself becomes a site of contestation and struggle. But this is a recursive process, Austin [2005] notes, since much of how we see ourselves is merely a reflection of the ways in which the rest of the world sees us. Identifying requires the simultaneous act of not identifying: to represent the self, one needs to project the other. This means that identity is in a constant state of flux and emergence and that it is not absolute. This fits the conclusions for this thesis and allows for a practical solution to the problem, which will be discussed at length in §6. Joseph [2012] defines identity as manifested in language in three ways: as the categories and labels that people attach to themselves and others to signal their belonging, as the indexed ways of speaking and behaving through which they perform their belonging, and as the interpretations that others make of those indices. Everyone has a set of combined national, ethnic, religious, generational and general identities. According to Joseph [2012], identities, whether of an individual or of a community, are not a given —they must forged textually and semiotically2. 2Through signs. 108 Chapter 4. Cultural Identity Language is the ultimate semiotic system, and every identity ideally wants its own language. For every identity, a time comes when a new generation wants it updated. New words, signs, and texts are admitted, often by hybridization with some more recent and vibrant identity (Joseph [2012]). Postmodernist theorists like Jacques Derrida and Michel Foucault concur with him in asserting that social practices are discursive since no social practice can take place outside the domain of meaning (Yin [2018]). Postmodernism does not do away with selves and identities but rather directs attention to how they are constructed, Howard [2000] cites. Achieving authenticity and meaning becomes a challenge when most human actors simultaneously experience a multiplicity of relationships and identities. She describes a conception of the self that is multidimensional and unified, emotional and cognitive, individual and social. Yet cultural identity poses the problem of how it is that even individuals who are not dissenters within their cultural identity nevertheless possess agency and make choices they perceive to be free while at the same time recognising that those choices are partly shaped by the culture of which they are a part (Joseph [2012]). This sense of cultural identity can, therefore, lead to classic determinism, which is often defended by enforcing a sense of authenticity (a very popular concept in the age of postmodernism), which leads to a full acceptance of one’s psychological and social location within a specific group of people, thus fully understanding oneself and one’s interpretation of life in terms of the particular civilization that contains them (Loughney [1998]). The notion of self has long been conceived as the autonomous, sovereign and sole locus or origin of experience, emotion and action in Western cultures (Yin [2018]). The individual’s autonomy is at the very centre of the conceptualisation and popular consciousness of personal identity. In previous chapters, we have repeatedly discussed the idea that true self-knowledge can only be gained through personal thought and experience. In these, the search and defence for an immutable self that controls outer behaviours across diverse situations becomes apparent throughout Western tradition. It has become the foundation of psychological literature, popular consciousness, and individual self-expression in Western cultures. However, this is not hegemonic. Hegel recognised that self-consciousness involves other human beings and is mediated through social value systems. Yin [2018] cites Carl Jung, a critic of Freudian psychoanalysis, as an 109 advocate of the self as a product of interactions with other human beings and the environment. In §4.1, we will encounter Hannah Arendt and her speech-based disclosure of identity. Many phenomenologists and cultural anthropologists alike side with this theory. Others have accepted the notion of the self as a product but stressed other sources. Karl Marx’s materialism, Yin [2018] writes, inverts the idealism in Hegel’s work by attributing identity production to the economic base. Nowadays, as the advancement of technology shrinks the world into a global village and brings people from different cultures into more contact, it paradoxically highlights cultural differences and social disparities and heightens the fundamental need for human connection beyond technological connectivity. Yin [2018] pushes non-Western peoples and cultures to go beyond Western paradigms and embrace their own cultural traditions in their question for selfunderstanding, self-definition, and self-assertion. Yin [2018] challenges Eurocentrism and Western domination in discourse, highlighting the necessity of communication in the realisation of the self. This way, she says, the world can form solidarities in their shared struggles over identity and self-determination. Individualism is the underlying assumption and defining characteristic of Western philosophy and modernity —the individual’s main concern is to keep their identity intact by separating from others and preserving boundaries at all costs. But setting borders inevitably leads to neighbourly disputes, which creates a rival mentality against others that, for instance, places individual liberty as a higher priority than moral order. Postmodernism wanted to eradicate the notion of the sovereign self in Western modern thought but is still very much drinking from it. Not only that, but it offers seemingly infinite options for individuals to distinguish and separate from others, acquiring new identities and creating universes of reality to satisfy their endless quest for meaning, identity and belonging. This hoists the individual’s desire and pursuit for authenticity, which brings us all the way back to the sovereign subject. According to Yin [2018], the focus on the individual has made it impossible to conceptualise a self-concept that is sensitive to social solidarity, communal good, equity and justice. It sustains and reinforces the traditional and modern Western individualist ontology that assumes the irreconcilability between the individual and the social. The assumption that the individual is prior to society is unique to Western cultures (Yin [2018]). This impedes a Western understanding of the non-Western cultures where the person is 110 Chapter 4. Cultural Identity thought to be born into a network of human relationships, and unity or holism is a concept central to the understanding of the self, human relations and social organisations —it assumes a totality (or cosmos) in which diverse and distinct human, natural and spiritual beings are interconnected and interrelated. The person does not define themselves by separating themselves from others but by being part of a network of relations in totality. Rather, the unique quality of the person is endowed by the uniqueness of the network of relations. This problem, however, has not gone over the heads of all Western thinkers. Loughney [1998] deems cultural identity in the sense that a person achieves the fullest humanity within an accepted context of traditional symbols, judgments, values, behaviour and relationships with specific others who self-consciously think of themselves as a community, a great contemporary challenge to many Western philosophical assertions about the person, society, meaning and truth. For non-Western cultures, historical memory provides a source of cultural identity, social cohesion, a sense of permanence amid change and a means of invigorating the present and shaping the future (Yin [2018]). This is being ripped apart by postmodernism, which has shattered history and memory into a non-continuous experience. Historical memory is also important in the Western world in that it shapes politics. In §4.2, the significance of a register of foundational events for contemporary culture will be explored in more detail. Cultural traditions are not defined in a fixed, ancient, or pure sense. No culture can exist without any influence from the outside in the long run. Cultural traditions are preserved and enriched through constant intracultural and intercultural communication (Yin [2018]). 4.1 Social and political identity Earlier in history, notes Howard [2000], identity was less of an issue; when societies were more structured, identity was largely assigned rather than chosen or adopted. In the present day, however, the concept of identity carries the full weight of the need for a sense of who one is, together with an often overwhelming pace of change in the surrounding social contexts. On that note, let us start this section with perhaps one of the most original contributions to the 20th century political thought. T¨ommel and d’Entreves [2024] make a compelling case for Hannah Arendt’s distinction of action (praxis) from fabrication (poiesis) by linking it to 4.1. Social and political identity 111 freedom and plurality, and her showing its connection to speech and remembrance, through which she articulates a conception of politics in which questions of meaning and identity are newly addressed. Arendt proposes three activities necessary to complete human life, each contributing in its distinctive way to the realisation of our capacities: labour, which caters to our biological needs of consumption and reproduction; work, which builds and maintains a world fit for human use, and action, which discloses the identity of the agent, affirms the reality of the world, and actualises our capacity for freedom —it is a mode of human togetherness. T¨ommel and d’Entreves [2024] observe that Arendt takes action to be that which distinguishes human beings from both the life of animals (who need to labour to sustain and reproduce themselves as well) and of the gods (with whom humans share, intermittently, the activity of contemplation). In the case of Autonomous Agents, Jackson [2017] remarks on the importance of utility. If the imitation human is not developed for practical use, we risk devolving into what he calls very anxious self-investigation—which might accurately describe the focus of this thesis. Questions quickly arise regarding the status, categorisation, and identification of Autonomous Agents as entities, as well as their legal, moral, sexual, ethical, and general responsibilities —especially if we deprive Artificial Intelligence of its role as a mere practical or profitable tool or as a scientific endeavour. The practicality of this camouflage, as Jackson [2017] puts it, has nonetheless been set aside in this work in order to address these challenging questions. According to Arendt, action is primarily symbolic: the web of human relationships is sustained by communicative interaction. Thus, speech articulates the meaning of actions and coordinates agents’ actions. Conversely, speech implies action, not only in the sense that speech itself is a form of action or that most acts are performed through speech, but that action is how we check the speaker’s sincerity (T¨ommel and d’Entreves [2024]). One of its central functions is the disclosure of the agent’s identity. In action and speech, Arendt maintains, individuals reveal themselves as unique individuals and reveal their distinct personalities to the world. One of the main drawbacks of action is its fragility, its susceptibility to the erosion of time and to forgetfulness. But remembrance alone, the retelling of deeds as stories, can save the lives and deeds of actors from oblivion and futility, thus enabling these acts to become sources of inspiration for the future —that is, models to be imitated, and, if possible, surpassed. This reflection we will 112 Chapter 4. Cultural Identity see mirrored in §4.2. The function of the storyteller is thus crucial not only for the preservation of the doings and sayings of actors but also for the full disclosure of the actor’s identity (T¨ommel and d’Entreves [2024]). Identity theory studies how individuals organise identity meanings, enact them in social situations, and respond to identity-relevant feedback (Davis et al. [2019]). Identities, which Davis et al. [2019] describe as internalised meanings attached to the self as a unique person, a role, and a group member, are placed into three overlapping bases: person, role, and group and social. Group identity refers to membership in a community, while social identity points to status categories, placing one within the broader social structure. T¨ommel and d’Entreves [2024] write that one of the crucial issues at stake in political discourse is the creation of a collective identity to appeal to when faced with the problem of choosing between alternative courses of action. Since there is always disagreement on that front, the identity of the community that is going to be created through each alternative becomes central. In this respect, political action and discourse are essential to the constitution of collective identities —a struggling process of constant renegotiation, in which actors articulate and defend competing conceptions of cultural and political identity. Arendt’s participatory conception of citizenship is particularly relevant in this context since it articulates the conditions for the establishment of collective identities (T¨ommel and d’Entreves [2024]). Once citizenship is viewed as the process of active deliberation about competing identities, its value resides in the possibility of establishing forms of collective identity that can be acknowledged, tested, and transformed in a discursive and democratic fashion. According to Arendt, one’s ethnic, religious, or racial identity was irrelevant to one’s identity as a citizen and should never be the basis of membership in a political community. Identity is a key mechanism through which social movements operate. In the words of Davis et al. [2019], advances in digital methods are particularly relevant to the social psychology of increasingly complex social movements —social media can provide a gateway to activism and collective identity formation. Social identity theory is concerned with identity processes through dynamics within and between groups, with attention to stereotyping, conflict, conformity, cohesion, leadership and organisational behaviour. Davis et al. [2019] see building bridges between identity theory and social identity theory as a key step in constructing a general theory of the 4.1. Social and political identity 113 self. Collective identity is a special case of social identity theory, which focuses on identity as it relates to social movements. Both identity theory and collective identity theory recognise social structures, situations, and networks as integral to identity processes. Social cognition and symbolic interaction are two of the traditional prevailing perspectives in sociological social psychology —a theory of how we store and process information. It has roots in psychology and the reliance on experimental laboratory methodologies, underlining social cognitive theories of identity (Howard [2000]). Cognitive schemas, which are abstract and organised packages of information, are the cognitive version of identities, and they can be of oneself or of groups (analogous to stereotypes, they include organised information about social positions and stratification statutes). They allow us to summarise and reduce information to key elements, but they also entail losing potentially valuable information. Moreover, they tend to be accompanied by means of qualitative categorisation and serve as explanatory services and justifications of social relationships —thus, social identities are embedded in sociopolitical contexts. Cognitive processes are also implicated in constructing, maintaining, and changing identities. Attribution processes (judgements of blame, causality, responsibility) are particularly relevant — and also biased based on intergroup identification and allegiances. Howard [2000] cites a theory by which knowledge structures would be collectively shared, thus increasing the emphasis on social processes and opening the door to a continuing recasting of social schemas as more flexible and more grounded in social interaction. This highlights the intertwinement of cognitive and interactional processes, which leads to the use of identity management strategies to manipulate group comparisons for purposes of social identification. Coleman and Williams [2013] define the self as a unique, identity-relevant set of attitudes, behaviours, and beliefs that provide information about what to do in enacting that identity, as well as specific discrete emotion profiles that, in turn, provide information about what to feel. Social identities are aspects of the self-concept derived from knowledge structures containing traits, attitudes, behaviours, and goals relevant to a social category. These identities provide coherence and allow people’s everyday actions to be understood and unified. Individuals, Coleman and Williams [2013] assert, may possess a constellation of discrete social identities. They emphasise the self that emerges 114 Chapter 4. Cultural Identity from group membership adoption of group characteristics into the self-concept, consistent with the social identity tradition, and specifically link social identities with emotion profiles. Davis et al. [2019]’s hypotheses, which are consistent with what identity theory predicts, and they later prove, is that identity feedback that deviates from the norm elicits negative emotion and active behavioural response. In contrast, identity verification evokes positive emotions and attenuated behavioural responses. Individuals seek to verify identity meanings through interaction, Davis et al. [2019] write. This process has four main components in a cybernetic feedback loop: the identity standard (the set of meanings attached to an identity), perceptual inputs (reflected appraisals), comparator (which contrasts the inputs against the identity standard), and behavioural outputs (and cognitive responses). Identity verification processes initiate when persons experience identity-relevant feedback. When reflected appraisals align with the identity standard, identity verification is achieved; if there is a significant distance, non-verification is experienced. Identity verification, according to Davis et al. [2019], elicits positive emotion, while non-verification brings distress —which is then intended to be minimized through cognitive and behavioural response: intensifying identity-relevant performances, altering cognitive perceptions of feedback, and when identity non-verification is persistent, changing identity meanings (Davis et al. [2019]). Howard [2000] asserts that symbolic interaction is founded on the premise that people attach symbolic meaning to objects, behaviours, themselves, and other people, and that they develop and transmit these meanings through interaction. Because meanings are developed through interaction, language plays a central role. The interactionist literature on identity formulates the construction, negotiation and communication of identity through language —at the most basic level, people actively produce identity through their talk. This, as we have seen, bears some resemblance to Arendt’s theory of identity. Language thus links the cognitive and interactive traditions (Howard [2000]). Identities locate a person in a social space by virtue of the relationships that these identities imply and are themselves symbols whose meanings vary across actors and situations. There are two main approaches: the concept of role identities, which focuses on the character forged in particular social positions, and identity construction and negotiation, which turns identities into strategic social constructions created through interaction. Howard 4.3. Art imitates life 121 4.3 Art imitates life Identity has been a recurrent topic in all human thought —not only academic, but artistic, too. As prefaced in §1.3, it has helped us separate us from the animals (albeit what animalists might say in §3.4) and other groups of human beings, up until we reached enough granularity in our assessment to know me from you or them. It is only fitting that this section begins with a legendary work which we already mentioned: Homer’s Odyssey (Homer [2019]). The ancient understanding of a person’s identity is glimpsed throughout the poem in a very particular custom shown repeatedly: to get to know a guest, they ask for their name, their birthplace, and who their parents were. But let us focus on the Odyssey’s protagonist. When Odysseus finally reaches Ithaca, he does so disguisedly: he wants to evaluate the state of his land before seizing it back, know allies from traitors. After a bloody massacre in which Penelope’s suitors and their accomplices are defeated, all that is left to do is reveal his true identity to his wife. But others have tried to trick Penelope before, and Odysseus has had a knack throughout his travels for inventing new identities for himself. One of them, when first reunited with Penelope. When she finds him after the battle, drenched in blood, they study each other. She recognises his face, but not his clothes, ratted; the castle’s nursemaid, Euriclea, had confided that the stranger before her bares the same scar on the leg her husband did from a childhood wound. But still, she doubts. Odysseus gets frustrated and asks Eurinome, the castle’s housekeeper, to prepare a bed in a different chamber than Penelope’s. Then, Penelope tries Odysseus one last time: she orders for her own nuptial bed to be moved to the room his supposed husband is asking for. Odysseus is horrified: he made that bed from a living olive tree and built the castle around it. It cannot be moved. Only then is Penelope convinced: this is something only Odysseus would know. And so, they are reunited. This scene, which takes place in Chant XXIII, the penultimate of the epic poem, presents us with a few examples of Homer’s (whomever, and however many, they might have been) notions of identity. We find that Penelope does not think bodily identity is enough: neither her husband’s features nor his telling scar (which was enough to convince Euriclea, his nursemaid) convinces 122 Chapter 4. Cultural Identity her of his identity. It is memory, or, more precisely, a very exclusive knowledge, that convinces her. We see different ideas on the vessel of one’s identity as we look at other cultures. For instance, in Ng˜ug˜ı wa Thiong’o 1980 Kikuyu language novel Devil on the Cross (wa Thiong’o [1980]), which focuses on politically challenging the role of international money and culture in Kenya, we encounter the idea that it is the heart which stores the soul of a person and thus, their identity. In the Kikuyu language, the word heart means many things: soul, spirit, conscience, mind, inner man, essence, etc. Back in the ancient Egyptian Kemetic philosophy, ib denoted both the heart and the mind, conscience and consciousness, Yin [2018] writes. It signified the human capacity of rationality and moral sensitivity. Though it did not flourish in the same continent, Confucianism shares similar ideas, Yin [2018] observes: the xin is the heart-mind and both conscience and consciousness, the affective and cognitive aspects of human awareness —and what makes one truly human. This goes to show that the heart as the nucleus of the self is a conception that has roots in more than one place in the non-Western world (Yin [2018]). In wa Thiong’o [1980], which I encourage the reader to read in full, the main characters discuss the profound ridges their country’s colonisation and post-independence corruption has created in their culture, philosophy, class struggles, nature... One of those issues is the matter of identity, which both involves the individual and the community they participate in. This reflects both Kemetic and Confucian ideas of the self, which are diametrically opposed to the dominant Western individualistic notion of the isolated, embedded, contingent and ultimately alone self, as we learn from Yin [2018] —instead, they see the self as a centre of relationships, the individual in relationship or the person in community. ’The human heart is flesh and it is not flesh. The heart makes a man and it is made by man. The heart is born by the body and in turn becomes the body. There is in man an organ called the heart. That organ is a kind of engine that pumps blood into the arteries and veins that carry food to all the cells of the body and remove the waste from all parts of the body. The organ co-operates together to make a human being see, touch, hear, smell, taste, talk, swing his arms, walk, start building his life. 4.3. Art imitates life 123 ’What he himself builds is the other heart. That other heart is the humanity we fashion with our hands, aided by our eyes, our ears, our noses, our mouths. That other heart is the product of our work and our actions, which are guided by our mind - the work and actions involved in modifying nature to make things to meet our needs, like shelter to keep out rain, clothes to keep out cold and sun, food to make the body grow and many other deeds. ’That humanity is in turn born of many hands working together, for [...] a single finger cannot kill a louse; a single log cannot make a fire last through the night; a single man, however strong, cannot build a bridge across a river; and many hands can lift a weight, however heavy. The unity of our seats is what makes us able to change the laws of nature and harness them to the needs of our lives instead of our lives remaining slaves of the laws of nature. [...] ’That humanity, fruit of the work of our hands and minds striving together to subdue nature, is what distinguishes a human being from a beast, a tree and all the other creatures in nature’s kingdom. [...] ’That humanity is the heart of a man because the heart of a man is linked irrevocably with the growth of his nature as a man. [...]’ (Ng˜ug˜ı wa Thiong’o in Devil on the cross) Kemetic thought and contemporary African understanding of personhood regard sociality or human relatedness as one of the most fundamental sources of self. According to Yin [2018], they are characterised by five dimensions: collectivity (self-understanding and self-development as a communal process defined by activities in and for the community), morality (to establish, preserve and enrich the common good in the community, society and humanity; it is central in the conception of the self), sensitivity (each individual must hold a reciprocal sensitivity and be responsive to others and the community: feelings are the basis of moral action), transformability (a person is teachable, malleable, capable of moral cultivation which leads to one’s higher self; this is inherent in the very existence of the human person) and inclusivity (all or none of humanity is chosen). In that regard, Hall [1990] writes: 124 Chapter 4. Cultural Identity We all write and speak from a particular place and time, from a history and a culture which is specific. (Stuart Hall in Cultural Identity & Diaspora) Virginia Woolf once said that fiction was likely to contain more truth than fact. Identity is not the exception to this trend, appearing more often than one would think. Brandon Sanderson’s The Rythm of War (Sanderson [2020]), the fourth instalment in an epic fantasy saga, paints it as an inevitable or passive force which compels everyone to do what is expected of them. The following fragment finds a divinity-adjacent creature which embodies the storm: “Take mercy upon them,” Dalinar said. “Temper your fury, Stormfather.” It is not fury. It is me. “Then protect them,” Dalinar said as the stormwall hit, plunging the illfated men into darkness. Should I protect all who venture out into me? “Yes.” Then do I stop being a storm, stop being me? “You can be a storm with mercy.” That defies the definition and soul of a storm, the Stormfather said. I must blow. I make this land exist. [. . . ] I am a storm. I cannot– You are not merely a storm!, Dalinar bellowed, his voice changing to rumbles of thunder. You are capable of choice! You hide from that, and in so doing, you are a COWARD! (Brandon Sanderson in The Rythm of War) It is interesting how elements of the earlier discussion can be discerned in this dialogue excerpt. First of all, it refers to the existence of an essence or soul of the self which determines the actions one can perform. So far we have discussed Cartesian dualism and soul theory as the 4.3. Art imitates life 125 two main schools of thought which argue for a somewhat intangible substance which anchors the self to an identity. At its core, the text questions the inevitability of what the sortal concept entails for the self —in this case, being a storm. Sortal concepts were particularly conspicuous in §2.2, although they kept making appearances while debating, as they did in §3.4 with the Abelardian predicates. Shortly after this exchange, a brief excerpt touches another facet of the discussion of identity: “Can we go inside?” Dalinar asked as they approached. You may, the Stormfather said. I cannot go inside [.. . ]. When a piece splits off, it is no longer me. (Brandon Sanderson in The Rythm of War) This seems to return us to the mereological essentialism introduced in §2.3 and §3: the understanding that an object cannot retain its identity if it loses any of its parts. But this is not exactly what the text aims for: it allows the larger self to retain its continuity, while the smaller fragment is stripped of it. This notion brings forth a host of logical problems, starting with the arbitrariness of choosing which part of the being retains its identity, which have been discussed throughout this thesis. Another fantasy story by Mark Lawrence, entitled The Book That Wouldn’t Burn (Lawrence [2023]), offers new thought on identity. The first fragment follows its protagonist as she studies a mechanical raven with an abrasive personality that serves as her guide through an infinite library: [...] were debating if the Raven was still the Raven or simply a collection of its pieces, devoid of the feisty, argumentative spirit that had previously animated [. . . ] (Mark Lawrence in The Book That Wouldn’t Burn) The fragment flirts with mereology and the possibility of an intangible relationship between the parts that makes the whole something else than its collection. It therefore teeters on the 126 Chapter 4. Cultural Identity edge of the debate about composition as identity (§2.3). This particular consideration echoes Aristotle’s Book Z in his Metaphysics (Aristotle [2018]), in which the Stagirite argues that the substance of a thing is its form or essence, rather than its matter or the composite of matter and form. He thus posits that form is constitutive of what the substance is in itself, not merely a property it happens to possess. It is in his assertion that the form inherent in the matter of which the substance is composed is the fundamental reality of the object, that Aristotle’s focus on finding the substrate or substance that binds all beings into a whole culminates. “Almost everything about you changes each time I meet you. Your clothes, your hair... your height... but if you ever stopped the flood of questions, I’d know it wasn’t you even if you looked exactly the same.” (Mark Lawrence in The Book That Wouldn’t Burn) This second fragment returns to the popular notion of psychological continuity —particularly, on trait over bodily continuity. The co-protagonist of the book is noting the prominence of the protagonist’s tendency to ask about everything, at all times. However, the comment is made in an endearing way and is not intended to reflect deeply on the protagonist’s identity. If it were, we would have to discuss the contingency of this character trait, and whether she would retain her identity if she stopped asking questions. The acclaimed film Everything Everywhere All at Once (2022) explores identity across worlds (§2.5). In it, both Evelyn (Michelle Yeoh) and her daughter Joy (Stephanie Hsu) jump through a multiverse of possibilities that branch out from every decision ever made, as well as divergent circumstances —such as a world in which everyone’s fingers are made of sausages, or another in which they are mere rocks. This leads to them meeting counterparts of their husband and father, Waymond (Ke Huy Quan), as well as experiencing the ways in which their own lives might have gone differently. Though a fantastic film touching a myriad of transcendental themes, it is particularly relevant for this thesis the fact that not once do they doubt the continuity of their identity as they jump from world to world, no matter how different their bodies, or histories, or surroundings might be. Theirs seems to be a very strong defence, whether intentional or not, of 4.3. Art imitates life 127 the notion of psychological continuity. The recurrence of this theory over all others may be due to its intuitiveness. An interesting contrast arises with the 2023 film Dream Scenario, which features Paul Matthews (Nicholas Cage), an unassuming college professor of evolutionary biology until, all of a sudden, people start dreaming of him —they might know him or they might have never seen him, but still he appears in their dreams with an unsure walk, aloof and out of place. Paul goes with it, blinded by his newfound fame. But the dreams shift to nightmares, with Paul’s dream version as the perpetrator of violent attacks against the dreamers, and the world begins to turn on him. No matter how much he tries to distance himself from the nightmare who bears his exact likeness, Paul is unable to separate himself from a perceived identity he had initially embraced. The film walks on a tightrope in matters of identity: when it first suits him, Paul identifies himself with his dream version, though he has no psychological or any other kind of continuity with it, aside for its likeness —which is not corporeal either. When he no longer can accept his dream version’s actions as his own, he rebuffs it; but it is too late, the world already perceives both versions as pertaining to one identity. The externality of identity allocation is central to the conclusions of this thesis (§6, §7), and will begin to come to the fore with the next chapter’s (§5) overview of legislation. In The Ten Percent Thief (Lakshminarayan [2020]) the issue of identity and identification in the Information Age5is directly addressed. The book transports us to a meritocratic corporate autocracy which manages its citizens by monitoring their perceived Productivity and keeping track of them through a comprehensive social network which universally identifies each individual. It becomes the BE-moji Project’s highest priority to fund a universally acceptable identity marker for each social media portal, and to devise a symbol for it. (Lavanya Lakshminarayan in The Ten Percent Thief ) The BE-moji project aspires to be the only true expression of one’s current state of mind, 5The Information Age refers to the current era, stretching from the mid-20th century onward, characterized by the increasing digitization of the world and the rapid shift from traditional industries to an economy centred on information technology. 128 Chapter 4. Cultural Identity communicating on one’s behalf at all times. Lakshminarayan [2020] creates a very interesting characterisation of its inception, where concerns about user privacy and control over their information are quickly dismissed in favour of what her characters perceive as a future without communicative restrictions: all is shared, all the time. This bypasses consent, of course, but is justified by the misguided ideal of a world in which one’s identity is known to all and integrated into a more efficient, overall better society. In doing that, the BE-moji Project becomes an example of identity being based on one’s consciousness, as well as external factors like one’s actions (see Arendt in §4.1) and social status. As expected, the manifestations of identity discussion in art are further muddied by the introduction of non-human actors. For the purpose of this work, we will focus on robotic entities and Artificial Intelligence. Several examples come to mind: from Her (2013), a film that follows Theodore, a lonely man who develops a romantic relationship with the female voice of an Artificial Intelligence model named Samantha, to 2015’s Chappie, about an Artificial Intelligence robot that is stolen and given new programming that allows it to think and feel like a human. Both challenge the definition of personhood and whether artificial minds might be considered hosts for autonomous identities, as well as brushing complicated topics like consciousness (§3.2). Among the most notable is Steven Spielberg’s A.I. Artificial Intelligence (2001), which is the story of David, a biocybernetic child capable of experiencing love. After being rejected by his assigned family, he embarks on a Pinocchio-esque journey to become a real boy. During his adventure, he first becomes aware of the chasm that separates biological robots like him from humans, only to lose his sense of individuality when he discovers that he is merely a copy of a mass-produced product. In the end, however, thousands of years after the events of the film, he is reactivated as the last witness of humanity and allowed to live for one day with a clone of his assigned mother. A few things stand out from this. The first is the conundrum it imposes to animalists (§3.4), with David being a sentient, half-organic being, which decidedly stresses the definition for animal. He may be an example of what the extinctionists are after (Rubin [2003]). David bears memory, the same body (though not completely organic, it experiences change). He has a consciousness, a sense of uniqueness which is shattered when he learns of his origins. On 4.3. Art imitates life 129 top of that, the familial relationship David first establishes with his assigned family involves him in their socio-cultural world; then, when he is rejected and left alone in the woods, he learns about the lives of other biocybernetic people, becoming an outsider to both worlds. His identity is fragmented, at odds with a lot of what this thesis has discussed so far. Then, to the troubles this film must cause for identity academics, we must add Monica’s clone, David’s assigned mother, whom he identifies as Monica herself, though this is a one-day lifespan clone of her, created some two thousand years after her death, and having Monica’s own memories and consciousness only to a degree. The focus on practicality we briefly mentioned in §4.1 appears in David’s creator’s initial discourse, in which he cites the reason for his creation as serving the human race in all the multiplicity of daily life. Another justification for his invention follows: To create an artificial being has been the dream of man since the birth of science. (Alan Hobby (William Hurt) in A.I. Artificial Intelligence) According to Bajema [2019], science fiction plays an important role in shaping our understanding of the implications of science and technology and helping us to cope with what the future holds. Harari [2024] remarks on the Great Robotic Rebellion of which the genre has warned us time, and time again. Gunkel [2023] claims it has already happened, albeit in a more mundane way than the march of robot armies. Bajema [2019] identifies three horrifying Artificial Intelligence monsters, though admits that the worst-case scenarios are still quite far: automation, supermachine and data. The automation monster foresees a future in which humans have been completely replaced by robots. It is a fear that stems from Czech playwright Karel Capek’s 1920 play Rossum’s Universal Robots, which coined the word robot for creations forced to perform the most undesirable activities and ends (spoiler alert) with a robot uprising and the extermination of most humanity. More widely known is Isaac Asimov’s I, Robot (1950), which predicts that humanity will eventually be unable to anticipate all the unintended consequences of rule-based systems and will lose control of robots as automated systems become more complex (Bajema [2019]). Stanley Kubrik’s HAL 9000, a supercomputer designed to automate most of the Discovery spaceship’s operations in 2001: A Space Odyssey (1968) is perhaps the most 130 Chapter 4. Cultural Identity known cautionary tale against seemingly foolproof automatization: after it makes a mistake, the crew decides to disconnect it, but HAL discovers the plot and manages to kill most of them. In WarGames (1983), a young hacker succeeds in launching a nuclear attack simulation in a supercomputer to which the military has given control of the nuclear attack plan. The situation escalates until the ensemble of the main characters succeed in teaching the computer the concept of mutually assured destruction, a scenario in which there are no winners. Bajema [2019] is very pessimistic towards an automated future: she predicts massive job loss, societal unrest, violent conflict. She then goes on to explore the terrifying moment when machine intelligence could match and surpass human intelligence, giving rise to supermachines. The Terminator (1984) and The Matrix (1999) are great, firmly secured in the pop culture sphere, expressions of that fear. The former presents a machine intelligence turned sentient called Skynet which identifies humanity as a threat in the near future and sets off to eliminate it. The latter imagines a future in which machines won the war and humanity was reduced to organic batteries. But Bajema [2019] fears the data monster the most: a three-headed beast composed of the Internet, digitalisation, and algorithms. She criticises the recent trend of political spin, outright lies, and allegations of fake news, which has only worsened since she wrote the cited chapter. She warns about monitoring, surveillance, and surrendering our own data —and how this leads to predictive control. The film (and the 1956 book of the same name by Philip K. Dick) Minority Report (2002) is a great example of such a practice taken to an extreme that now seems less and less far-fetched: a world in which crime is predicted and not responded to, so that criminals are arrested before they commit any crime. This, as the story goes, is a flawed system —we need only look at a scenario that falls closer to everyday life, with online shops scrambling to predict our every need just from previous activity. We might all agree that most of the time this automated curation misses the mark. Still, Bajema [2019] is still hopeful: she predicts that human adaptability, ingenuity, and resilience may be able to slay them before they grow too much in power. The following chapter visits one of the longest-standing battlefronts in this struggle: regulation. 5.1. Legislative triangulation 137 The law, known as Unlawful Production Or Distribution Of Certain Sexually Explicit Videos, specifies the scope of the penalty it regulates by stating: A person commits an offence if, without the effective consent of the person appearing to be depicted, the person knowingly produces or distributes by electronic means a deepfake video that appears to depict the person with the person’s intimate parts exposed or engaged in sexual conduct. (Tex. Pen. Code §21.165) At the federal level, the Deepfake Task Force Act (Homeland Security and Governmental Affairs, Senate: [2021]) and other proposed regulations reflect growing concern about the national security implications of deepfakes. China has taken a unique approach in addressing the challenges of the Information Age: it has implemented a comprehensive approach to digital identity through their Real-Name Registration laws (Lee and Liu [2016]), which require individuals to link their online activities, including social media accounts and mobile services, to their legal identities. Although this greatly reduces online anonymity, China is prioritising collective security and social harmony over individual privacy in the digital realm. This approach is consistent with China’s broader governance philosophy, which often places societal interests above individual ones. The digital identity framework is part of a broader strategy that includes the Social Credit System (Kobie [2019]), which further integrates personal data into state governance. The system aims to promote trustworthiness and integrity in society by incentivising positive behaviours and discouraging negative ones. China has implemented a comprehensive approach to digital identity through its Real-Name Registration laws (Lee and Liu [2016]), which require individuals to link their online activities, including social media accounts and mobile services, to their legal identities. It reflects China’s prioritization of social stability and national security in its governance model: the government significantly reduces online anonymity, emphasizing collective security and social harmony over individual privacy in the digital sphere. This approach aligns with China’s broader governance philosophy, which often prioritizes societal interests over individual rights. 138 Chapter 5. Legal Identity The digital identity framework is part of a larger strategy that includes the Social Credit System (Kobie [2019]), which further integrates personal data into state governance. This system aims to promote trustworthiness and integrity in society by incentivizing positive behaviours and discouraging negative ones, reflecting China’s emphasis on social order and collective responsibility. This approach to digital governance represents China’s unique path in addressing the challenges of the internet age, balancing technological advancement with social stability in a way that reflects its cultural and political context. However, in 2021 the Personal Information Protection Law of the People’s Republic of China (PIPL) (People’s Republic of China [2021]) was enacted. It granted every citizen the right to manage their own personal information, with clauses guaranteeing the right to add, amend or delete their own data on their own initiative. It also contains provisions on the right to be informed by organisations about the use of one’s data and the transfer of the right of access after the death of a natural person. T he article most relevant to this thesis, as presented in People’s Republic of China [2021], states: An individual has the right to know and make decisions about the processing of his/her personal information and the right to restrict or refuse others’ processing of his/her personal information unless otherwise provided for by laws and administrative regulations. (Personal Information Protection Law of the People’s Republic of China, Article 44) In the General Principles of Civil Law of the People’s Republic of China People’s Republic of China [2017], a natural person is defined as a citizen —a person or individual. This is in line with the EU and US interpretations of the term. A distinction is made, as in other regulations, from a legal person, which is more precisely defined: Article 36. Legal persons are organizations with civil capacity, are competent to perform civil acts, and, according to law, independently enjoy civil rights and assume civil duties. 5.1. Legislative triangulation 139 The civil capacity of a legal person and its competence to perform civil acts arise when the legal person is established and cease when the legal person terminates. Article 37. A legal person must satisfy the following conditions: 1. Be established in accordance with law; 2. Possess the necessary property or funds; 3. Possess its own name, organizational structure, and premises; 4. Be able to assume civil obligations independently. (General Principles of Civil Law of the People’s Republic of China, Chapter 3, Section 1) From this, one can infer that legal persons are corporations and other non-human, human-run entities. Only natural persons are included under the umbrella of online real-name registration policies, which use various methods: e.g., government regulation of Internet Service Providers (ISPs) to technologies embedded in communication infrastructure, or tracking the identity of the speaker. On the other hand, both natural and legal persons are observed in the implementation of the social credit system, which is a mix of initiatives aimed at regulating the financial credit industry, easing data sharing between government agencies, and promoting state-sanctioned moral values, although the latter goal remains somewhat vague (Yang [2022]). Financial creditworthiness is a familiar concept in Western societies —it documents the financial history of individuals or companies and predicts their ability to repay future loans. Because China’s market economy is relatively young, the country lacks a reliable system for accessing the financial records of other individuals and companies. Developing such a system is a vital and largely uncontroversial goal. The road to deciding what can be included in the social credit system has not always been straightforward, sometimes mixing the two concepts. According to Yiqing [2019], in 2019, the National Health Commission (NHS) published a document encouraging local authorities to explore the inclusion of voluntary blood donation. It was criticised nationwide, showing that the 140 Chapter 5. Legal Identity country seems to prefer to keep both domains mostly separate —as evidenced by the draft law published in 2022 (Creemers [2014]), which treats them with different sets of rules. However, some local governments have attempted to blur the lines —for example, in August 2022, the Liaoning province announced it was nevertheless exploring how to reward blood donations within the financial credit system (Canales and Mok [2022]). Although the government has largely resolved the financial aspect, regulations on the social credit side have been inconsistent and ambiguous. To date, a system has been developed that focuses primarily on enterprises rather than individuals —it aggregates data on corporate compliance from several government agencies. Its output is available to any Chinese citizen through a portal containing records on individuals and other organisations. The portal serves as a central hub for more than three dozen (often highly specific) databases, including lists of people who have failed to pay court judgments, legitimate Chinese universities, companies approved to build robots, and hospitals found guilty of insurance fraud. Yet the compilation seems so arbitrary that it is hard to see how people could use the portal as a comprehensive data source. China’s approach to identity theft is intertwined with its tight control over digital identities. The Real-Name Registration system reduces the risk of identity theft by linking all online activity to a verified identity. However, the centralised nature of this system means that any breach or misuse of data by the state or hackers could have serious consequences. The government has enacted several Articles in its Criminal Law of the People’s Republic of China (People’s Republic of China [1997]) that severely punish identity theft with jail time and hefty fines, illustrating the government’s strong emphasis on maintaining control and security in the digital world (Ning and Wu [2023]). One of them follows: Whoever forges, alters, buys, sells, steals, forcibly seizes, or destroys the official documents, certificates, or seals of a State organ shall be sentenced to fixed-term imprisonment of not more than three years, criminal detention, public surveillance, or deprivation of political rights; if the circumstances are serious, he shall be sentenced to fixed-term imprisonment of not less than three years but not more than 10 years. (General Principles of Civil Law of the People’s Republic of China, Article 5.2. The net of identity legislation 141 280) China has passed laws that explicitly prohibit the creation and distribution of deepfake content without consent. The government is particularly concerned about their potential to disrupt social stability and has integrated the issue into its broader Internet governance strategy, which aims to control it within its borders (Huang and Kurlantzick [2020]). The Cyberspace Administration of China (CAC, Lee [2022]), together with the Ministry of Industry and Information Technology and the Ministry of Public Security, adopted the Provisions on the Administration of Deep Synthesis Internet Information Services in 2022 (Translate [2022]), which took effect in 2023. The regulations require deepfake (also referred to as deep synthesis) content to be clearly labelled as such. Severe penalties are imposed for using it to produce, copy, publish or disseminate information banned by laws and regulations or for failing to comply with user registration and identity verification systems, among other things (Finlayson-Brown and Ng [2023]). The rules apply to individuals who use deepfakes and legal entities or platforms that host such content without proper data management and disclaimers. 5.2 The net of identity legislation The notion of personality rights is central to the upcoming discussion. According to Neethling [2005], these recognise a person as a physical and spiritual-moral being and guarantee their enjoyment of their own sense of existence. Neethling [2005] lists the specific personality rights he thinks should be granted. The list is headed by the right to life, a person’s more valuable asset; others include the right to physical liberty or privacy, which will be particularly important in this section. The last one is the right to identity, which is infringed when any of these facets are used in a way unreconcilable with one’s true identity, through falsification or misidentification. The word privacy, of which both Hill [2024] and Kulhari [2018] note the difficulty in defining it, was most famously described in a Harvard Law Review article in 1890 by lawyers Samuel D. Warren II and Louis D. Brandeis (who would eventually join the Supreme Court). Their proposal was succinct, and it called for the right to privacy to be protected by law in a way that has not materialised to this day. To them, privacy was: 142 Chapter 5. Legal Identity [T]he right to be let alone. (Samuel D. Warren II and Louis D. Brandeis in The Right to Privacy) ¨ Unver [2018], on the other hand, defines it as: [F]reedom from unauthorised intrusion. (H. Akin ¨ Unver in Politics of Digital Surveillance, National Security and Privacy) Roger Clarke forgoes the monolithic nature of privacy in favour of distinguishing four categories: of the person, of behaviour, of data, and of communication; thus allowing each kind of privacy to be protected differently (Kulhari [2018]). His thesis discusses protecting personal data in the context of ensuring privacy, suggesting a natural (though controverted) link between the right to privacy and the right to data protection. Kulhari [2018] characterises one facet of its polemic nature through the EU’s ambiguous decision of rooting the right to privacy in its Charter of Fundamental Rights while seemingly treating data protection as an economic matter. But to try and differentiate between the two on these grounds, she argues, is but a red herring. Another attempt to sever one from the other comes from the way the GDPR (European Union [2016]) is worded: it does not reference the right to privacy, but Kulhari [2018] maintains that the right to data protection is nuanced and builds on the premise that data processing is inadvertent. Accordingly, it follows that the GDPR contains detailed provisions regarding the obligations of the data controller and processor. The emphasis on how and when personal data can be legitimately processed is a corollary to the right to protect one’s personal data. Identity is inevitably linked to the protection of privacy the moment it is distinguished from identification. Neethling [2005] notes the difference in identifying a person’s name, image, or any other facet of their identity with true personal facts. Those personal facts are part of the private facts to protect, while conversely steadily becoming an essential enable of today’s digital society, Kulhari [2018] writes. She observes that the GDPR (European Union [2016]) does not include a distinct right to identity, so safeguarding identity requires the exercise of the 5.2. The net of identity legislation 143 available tools within the concept of lawful processing. The issue worsens, Kulhari [2018] goes on, with attributes such as pertinence to a group. Then, the profile is not inferred solely from the personal data of the profiled person but uses a large amount of data relating to many other people, which may or may not be anonymised. Moreover, the process results in attributing certain characteristics to an individual derived from the probability that they belong to a group and not from data communicated by them. When Neethling [2005] was written, the right to identity had been recognised in countries such as Italy, France, Switzerland and South Africa. In the US, the crimes of public defamation and appropriation of the defendant’s name or likeness are examples of the protection of identity rather than privacy (The State of California [2019], The State of Texas [2023], Homeland Security and Governmental Affairs, Senate: [2021]). Similarly, in Germany, the offences of falsifying or distorting a person’s image and the economic misappropriation of a person’s name or likeness are sanctioned. Still, the right to identity has yet to be generally recognised by the courts. In England, identity is not explicitly protected but may be implicit in other offences such as malicious falsehood and passing off. Although publication of falsehood is, in principle, unlawful, infringement of identity may be justified by freedom of expression, as with defamation. In 2024, fifteen minors were condemned to a year of probation and gender equality and social-affective courses in Badajoz (Spain) for creating and distributing artificially generated pictures of some of their female classmates or acquaintances (Ricou [2024]). They were collectively charged with twenty counts of child pornography and against moral integrity. They had been shortly preceded by the one-year incarceration of an eighteen-year-old for doing the same to two of his high-school classmates in Navarra (Spain). Neethling [2005] notes that in, for example, Germany or France, the rights to a name, image, and written and spoken word are regarded as separate personality rights. In 2007, South Korea became the first nation to implement real-name Internet. The system required users to verify their identities by submitting their Resident Registration Numbers (RRNs) before posting comments or uploading content on certain websites (Caragliano [2013]). In 2020, Naver, a South Korean online platform which offered, among other products, a search engine, email hosting, blogs, maps, and mobile payment, decided to disclose the online comment 144 Chapter 5. Legal Identity histories of all its users (Lee et al. [2024]). Their goal was to alleviate the social problems caused by malicious online comments on their news website by allowing anyone to view the comment history of other users. The announcement was issued from one day to the next, so users did not have sufficient time to find and delete any comments they did not want to be shown to the public. Lee et al. [2024] explains that the new system disclosed a suer’s comment history, including their screened ID, nickname, list of comments and the articles on which they were posted, reactions of other users to recent comments (e.g., likes and dislikes), and the comment-management measures (e.g., number of deleted comments). However, in 2021, the law was ruled unconstitutional and abolished, with lawmakers arguing it undermined the freedom of anonymous expression and the users’ right of self-determination in disclosing personal information (Lee et al. [2024]). In 2024, the National Assembly in South Korea passed the Act on the Development of Artificial Intelligence and Establishment of Trust (AI Basic Act). It is the second national regulation established to govern this matter, following the European AI Regulation. The Act includes various measures to support the development of AI, such as promoting research and development, facilitating access to academic data, and establishing an artificial intelligence data centre. In 2021, a China-Africa initiative to pursue “common development and security, jointly building a community with a shared future in cyberspace” (Cyberspace Administration of China [2021]) was launched. In it, China called on Chinese and African governments, Internet companies, technical communities, social organisations and individual citizens to work together to build a cyberspace for common development —upholding the philosophy of “achieving shared development, ensuring common security, realizing joint governance, and enjoying benefits together” (Cyberspace Administration of China [2021]). Some of the proposed points included improving Internet access and promoting connectivity, advancing information infrastructure construction, strengthening strategic mutual trust in cyberspace, improving data security management and protection of personal information, reinforcing international cooperation to combat cybercrime and cyberterrorism, promoting capacity building on cyberspace governance, strengthening support and assistance to vulnerable groups and leaving no one behind, and contributing to the implementation of the 2030 Agenda for Sustainable Development (United Nations - Department 5.2. The net of identity legislation 145 of Economic and Social Affairs [2015]). China had high hopes for the initiative, closing its statement by saying that “[t]he Internet is a shared home for mankind” Cyberspace Administration of China [2021] and reaffirming its willingness to work with African countries to seize opportunities and meet challenges “to jointly build a closer community with a shared future in cyberspace for a better future for mankind” (Cyberspace Administration of China [2021]). For its part, Africa studied cyberspace’s rise for many years before the China-Africa initiative. As early as 2000, the African Union’s Convention on Cyber Security and the Protection of Personal Data discussed the threat of cybercrime and drew up a document, African Union [2014], to establish mechanisms to combat it, starting with the strengthening and modernisation of criminal law, while seeking “a balance between the use of information and communication technologies and the protection of the privacy of citizens in their daily or professional lives, while guaranteeing the free flow of information” (African Union [2014]). From there, each country has followed its own trajectory towards 2025. In 2013, the South African Parliament passed the Protection of Personal Information Act (POPIA) (Parliament [2013, 2020]), which came into force in 2020. It covers a very broad spectrum of data management, from regulating big data to banning the collection of children’s digital information, including issues as diverse as cookies, consent, limitations on the rules governing the use of information in the media, and the right to be forgotten. It is overseen by the Information Regulator of South Africa, an independent body empowered, among other duties, to monitor and enforce compliance by public and private entities ((South Africa), Information Regulator [2013]). But South Africa is not free of the serious mistakes that can be made with the introduction of new technologies: by 2022, alarms were sounded as surveillance companies outnumbered police stations by more than 10 to 1, installing cameras that could be equipped with facial recognition software under the umbrella of private organisations that would profile and track citizens without official jurisdiction (Hao and Swartarchive [2022]). More harshly, Zimbabwe’s president since 2017, Emmerson Mnangagwa, unveiled in 2018 an ambitious Smart Cities initiative he assured would deliver digitally connected cities that would transform people’s lives and rebuild roads, bridges, schools, clinics and homes (Ndlela [2020]). The modernisation project would require 146 Chapter 5. Legal Identity “high-level security systems”, the result of a partnership with China’s Cloudwalk Technology Co, an organisation tasked with developing a facial recognition system that would see its first stone laid when Huawei installed a network of public surveillance cameras in 2020 (Ragazzi et al. [2020]). Perhaps more worryingly, the Zimbabwean government had to hand over large amounts of photographic data to Cloudwalk so that the software could be tailored to the Black population. This raised concerns about privacy and data protection. Mnangagwa’s government has been accused of drifting towards authoritarianism, Ndlela [2020] reports, and now the technology for digital surveillance is being increasingly perfected, nearing the point where it can be used to undermine free speech and political opposition. While the government claimed that surveillance cameras are common in several countries and have been used to combat crime and promote security in public spaces and on roads, human rights defenders argued that Zimbabwe does not have robust privacy and data protection laws to prevent abuse or misuse of the data collected by facial recognition cameras. Some, among which stand Ndlela [2020] and of Southern Africa [MISA], have raised concerns about the choice of China for this partnership, while others (Ragazzi et al. [2020]) have voiced reservations about the transfer of facial data from government hands to a private company. This highlights the current lack of laws regulating the cross-border transfer of Zimbabweans’ data from Zimbabwe to other countries, although there are sections in the Zimbabwean Constitution that protect citizens’ personal information (of Southern Africa [MISA]). Crossing the Atlantic, we find ourselves in South America. In this section, we will discuss Brazil and its ban on EU-based Meta from training its AI models on Brazilian personal data (Weatherbed [2024]). Suffice it to say this is not the only example of data protection policies being enforced in the region. In 2024, Brazil’s National Data Protection Authority (ANPD) cited “risks of serious damage and difficulties for users” (Weatherbed [2024]) in Meta’s privacy policy update, in which the company gave itself permission to use public Facebook, Messenger and Instagram data for AI training. It had been pointed out that one of the largest image caption datasets used for such purposes contained personal, identifiable photos of Brazilian children, putting them at unacceptable risk of being used for deepfakes and other non-consensual uses of