​

The Autonomous Mind: The Right to Freedom of Thought in the Twenty-First Century

​

• Department of Psychiatry, Trinity College Dublin, Dublin, Ireland

To lose freedom of thought (FoT) is to lose our dignity, our democracy and our very selves. Accordingly, the right to FoT receives absolute protection under international human rights law. However, this foundational right has been neither significantly developed nor often utilized. The contours of this right urgently need to be defined due to twenty-first century threats to FoT posed by new technologies. As such, this paper draws on law and psychology to consider what the right to FoT should be in the twenty-first century.

 

After discussing contemporary threats to FoT, and recent developments in our understanding of thought that can inform the development of the right, this paper considers three elements of the right; the rights not to reveal one's thoughts, not to be penalized for one's thoughts, and not to have one's thoughts manipulated. The paper then considers, for each element, why it should exist, how the law currently treats it, and challenges that will shape it going forward. The paper concludes that the law should develop the right to FoT with the clear understanding that what this aims to secure is mental autonomy. This process should hence begin by establishing the core mental processes that enable mental autonomy, such as attentional and cognitive agency.

 

The paper argues that the domain of the right to FoT should be extended to include external actions that are arguably constitutive of thought, including internet searches and diaries, hence shielding them with absolute protection. It is stressed that law must protect us from threats to FoT from both states and corporations, with governments needing to act under the positive aspect of the right to ensure societies are structured to facilitate mental autonomy. It is suggested that in order to support mental autonomy, information should be provided in autonomy-supportive contexts and friction introduced into decision making processes to facilitate second-order thought. The need for public debate about how society wishes to balance risk and mental autonomy is highlighted, and the question is raised as to whether the importance attached to thought has changed in our culture. The urgency of defending FoT is re-iterated.

“Carefully guard your thoughts because they are the source of true life.”

​

Proverbs 4:23

“[I]t is proper to take alarm at the first experiment on our liberties… The freemen of America did not wait till usurped power had strengthened itself by exercise, and entangled the question in precedents. They saw all the consequences in the principle, and they avoided the consequences by denying the principle.”

James Madison (1785)

​

To lose sovereignty over our minds is to lose our dignity, our democracy, and even our very selves. Such sovereignty is termed mental autonomy. This is “the specific ability to control one's own mental functions,” which include attention, memory, planning, rational thought and decision making (Metzinger, 2013). Dignity, “the presumption that one is a person whose actions, thoughts and concerns are worthy of intrinsic respect, because they have been chosen, organized and guided” (Nuffield Council on Bioethics, 2002, p. 121, italics added) requires mental autonomy. Democracies, in which citizens choose the laws that bind them (Johnson and Cureton, 2019), are only possible if citizens are mentally autonomous. The ability to think freely is so essential to our identity that to violate it is to deprive us “of personhood altogether” (Halliburton, 2009, p. 868).

​

Law must hence safeguard mental autonomy. International human rights law does this through the right to freedom of thought (FoT) (Nowak, 1993). In the United States, FoT is protected by the Bill of Rights (Blitz, 2010; Richards, 2015). Given the centrality of the right to FoT to both personhood and democracy, one may expect it to be clearly defined and frequently exercised. One would be wrong in both cases.

​

The right to FoT is poorly defined. Attempts to sketch its contours have been negligible (Kolber, 2016). It is unclear what counts as thought, what qualifies as a violation of the right, whether the right should be absolute and, if not, what would justify its violation (Mendlow, 2018). As a result, the right is rarely invoked. In the United States its exercise most often involves cases of sexual thoughts about minors (e.g., U.S. v. Gamache, 1998; Doe v. City of Lafayette, Indiana, 2003; U.S. v. Bredimus, 2003; U.S. v. Kaechele, 2006; U.S. v. Tykarsky, 2006; U.S. v. Stokes, 2013; for a related case see State of Washington v. Stevenson, 2005). Has the right to FoT, a foundational human right, really degenerated into merely the right to think the unspeakable?

​

Nineteenth-century technological advances spurred legal resistance that yielded the new conception of a right to mental privacy (Warren and Brandeis, 1890). Today, twenty-first century technological advances pose new threats to FoT. These demand we clearly draw the contours of the right to FoT to ensure our mental autonomy in this new landscape. This process needs to be informed by developments in the psychological understanding of thought. As such, this paper will consider, in an interdisciplinary manner, what the right to FoT should be in the twenty-first century.

​

Challenge I: New Technological Abilities to Access Thought

Behavior-Reading

Humans bleed data. For the longest time, it seeped into the earth where it fell or was washed away by the tides of time. Now, new technologies capture and store it indefinitely. Technology behemoths, such as Facebook and Google, possess unprecedented troves of user data. Data-analytic companies hold thousands of data points on millions of people (e.g., i360, 2019). National security agencies can access even greater volumes of data (Greenwald, 2013).

​

This data can now be analyzed, using advanced machine learning algorithms, to infer unobservable inner states of individuals. This can be termed “behavior-reading.” People's observable behaviors, including their facial expressions, actions, possessions, purchases, musical preferences, websites visited, words used in Facebook posts, and “likes” registered on social media, can be used to make inferences about their unobservable inner world (Rentfrow and Gosling, 2003; Golbeck et al., 2011; Kosinski et al., 2013; Wang and Kosinski, 2018). For example, deep learning neural networks are now better at detecting people's sexuality from their faces than other humans are, posing a “threat to the privacy and safety of gay men and women” (Wang and Kosinski, 2018, p. 246). Social network users have been shocked by reports that detailed personal information about them could be predicted from simply what pages on Facebook they had “liked” (Kosinski et al., 2013).

​

This is utilized for financial and political ends. Surveillance capitalism claims human experience as free raw material that can be used to deduce “thoughts, feelings, intentions, and interests,” predict behavior, and then be monetized by selling such predictions to advertisers (Zuboff, 2019, p. 81). Micro-targeting uses personality data gathered on individuals to target them with political adverts which should, in theory, be maximally effective. This was at the heart of the Cambridge Analytica scandal (Cadwalladr, 2018). Whilst the effect of micro-targeting in changing political views is probably small (Hersh and Schaffner, 2013; Liberini et al., 2018; see also Kalla and Broockman, 2018), it is likely to have a greater effect on turnout (Bond et al., 2012). In close political contests, this could have decisive effects.

​

Whilst it is always difficult to separate political bluster and marketing spiels from fact, the use of behavior-reading poses a potential threat to the FoT of technology users. A century ago, U.S. Supreme Court Justice Brandeis crusaded against big banks use of “other people's money” (Rosen, 2016). Today, we need to address big tech's use of other people's data.

Brain-Reading

Whilst care needs to be taken to avoid overstating the ability of neuroscience to decode thoughts from people's brain activity (“brain-reading”), there has been significant progress in this area. Whilst some work has studied how political stances can be predicted from brain structure (Kanai et al., 2011) and function (Schreiber et al., 2013), the majority of research has focused on decoding specific thoughts and perceptions.

​

Early studies were able to predict with a high degree of accuracy what type of pre-defined object (face, house, cat, etc.) an individual was looking at, based on their neural activity (Haxby et al., 2001). Technology can now use people's neural activity to predict what novel images people are viewing (Kay et al., 2008; Naselaris et al., 2009). Neuroimaging could once only create rudimental reconstructions of what an individual was seeing (Miyawaki et al., 2008). Now it can create “remarkable” reconstructions (Nishimoto et al., 2011). Brain imaging data can also be used to decode what people are hearing (Mirkovic et al., 2015) and predict what objects people have dreamt of (Horikawa et al., 2013).

​

Research has also attempted to infer individuals' verbal thoughts from their neural (Martin et al., 2014; Wang et al., 2017) and throat musculature activity (Jorgensen et al., 2003) with some success. Neural activity when people mime speech can now be directly converted into recognizable speech (Anumanchipalli et al., 2019). A major advance has been the ability to decode thoughts individuals are having which the decoding system has not previously been trained to recognize (Anderson et al., 2016; Pereira et al., 2018). Such approaches work on the principle that words are represented in the brain as semantic vectors. That is, any word can be represented by the extent it is associated with a finite number of specific features (Anderson et al., 2016). Each word in a basic 30,000 word vocabulary can be represented by a unique pattern of scores on these vectors. Once you know the neural activity associated with each of these features, you can potentially decode any word someone thinks.

​

Basic intentions can be decoded too (Haynes et al., 2007). It is even possible to predict what participants will do (for very basic tasks) before they themselves know (Soon et al., 2008). The ability to use this information to predict people's intentions in real-time is still limited. For example, the ability to predict people's intentions from their brain activity whilst they are playing a first-person shooter computer game in a scanner is currently limited to predicting intended movements. Intentions to fire cannot be detected due to the neural signal being swamped by a surge of activation associated with the emotions of firing or being fired upon (Smith, 2013).

​

Much of this work is preliminary, with accurate predictions only being possible under highly constrained experimental conditions. Yet funding is pouring in. Governmental funding in the United States has come from the Army, Air Force, Intelligence Advanced Research Projects Activity (IARPA) and Defense Advanced Research Projects Agency (DARPA) (Martin et al., 2014; Wang et al., 2017; Yuste et al., 2017), suggesting national security interests in brain-reading. Significant private investment is also occurring. Annual spending on neurotechnology by for-profit industry is estimated to be US$100 million per year and growing (Yuste et al., 2017). Facebook has announced plans for a Brain Computer Interface that aims to decode users' thoughts and transmit them to Facebook (Solon, 2017), and is actively funding research in this area

(Anumanchipalli et al., 2019). Microsoft has patented brain-reading technology (Keskin et al., 2018). Elon Musk's start-up, Neuralink, is also trying to develop brain-computer interfaces (Marsh, 2018). Progress could be rapid and have profound implications for FoT.

Taking Action Early: The Precautionary Principle

The law has long sensed the potential for technological advances to threaten FoT. “Advances in the psychic and related sciences,” prophesized U.S. Supreme Court Justice Brandeis in 1928, “may bring means of exploring unexpressed beliefs, thoughts and emotions” (Olmstead v. United States, 1928). In the 1970s, U.S. Supreme Court Justice Brennan noted that, “[t]he central storage and easy accessibility of computerized data vastly increase the potential for abuse of that information, and I am not prepared to say that future developments will not demonstrate the necessity of some curb on such technology” (Whalen v. Roe, 1977).

​

Such threats have now either materialized or are imminent (Hanson et al., 2019). The Morningside Group (a collaboration of neuroscientists, neuro-technologists, clinicians, ethicists, and machine-intelligence engineers) recently observed that “technological developments mean that we are on a path to a world in which it will be possible to decode people's mental processes and directly manipulate the brain mechanisms underlying their intentions, emotions and decisions” (Yuste et al., 2017). As such, they argue that it is “crucial to consider the possible ramifications now” (p. 160). Considering the implications of such technological developments for the right to FoT is one such crucial task.

​

A clear justification for this is offered by the precautionary principle. The World Commission on the Ethics of Scientific Knowledge and Technology (COMEST) defines the precautionary principle as: “When human activities may lead to morally unacceptable harm that is scientifically plausible but uncertain, actions shall be taken to avoid or diminish that harm.” (COMEST, 2005, p. 14). It defines actions as “interventions that are undertaken before harm occurs that seek to avoid or diminish the harm.” As Alegre (2017) has argued, technological developments that have the capacity to interfere with our FoT fall “clearly within the scope of ‘morally unacceptable harm”' (p. 230).

​

Challenge II: New Understandings of Thought in the Twenty-First Century

What Is Thought?

One problem for developing the right to FoT is that no definition of “thought” has been provided. As Loucaides (2012) notes, “there is no adequate material in the preparatory works of the drafters of the European Convention regarding the concept of ‘thought”' (p. 80). When studying the mind, psychologists tend not to use the blanket concept of “thought.” They work with more clearly specified mental processes, including memory, attention, inner speech, mental imagery, decision making, and planning (Alderson-Day and Fernyhough, 2015). Philosophical investigations of mental autonomy mirror this (Metzinger, 2013). As the purpose of FoT is to facilitate mentally autonomous individuals, discussions of the right to FoT need to track this approach.

​

It will hence be important to identify the key elements of thought that enable mental autonomy. Metzinger (2013) has proposed these are attentional agency and cognitive agency. Attentional agency is the ability to control one's focus of attention. This ability is under increasing threat as social media uses insights from the behavioral sciences to design products that more and more effectively hijack and retain attention (Eyal, 2014; Pandey, 2017), to the point of addiction (Kuss and Griffiths, 2017). The negative effects of this can be specific. The hijacking of parent's attention by smartphones can have detrimental effects on their families (Kildare and Middlemiss, 2017). However, there is also a wider problem; the person who cannot control their attention cannot control their thoughts. And, as Metzinger (2015) notes, “for as long as one cannot control one's own thought one cannot count as a rational individual” (p. 272).

​

Attentional agency is needed for cognitive agency, the ability to control goal/task-related, deliberate thought (Metzinger, 2013). An essential element of cognitive agency is the second-order mental action. To understand what this is, it is necessary to consider “hierarchical” accounts of autonomy (e.g., Frankfurt, 1971). These begin with the thoughts, desires and impulses that well up within us, termed “first-order” mental actions. To act on these, without reflection, is to fail display autonomy. Frankfurt (1971) called creatures who only acted in this way “wanton.” In his view, they were “not persons” (p. 11). An alternative response to first-order mental actions is to perform more mental work (a “second-order” mental action) to determine if such thoughts/desires/impulses are authentic, i.e., consistent with one's own chosen values and goals. Second-order mental actions make thoughts and desires “more truly one's own,” and allow people to recognize that their first-order thoughts may represent “a force other than” their own (Frankfurt, 1971, p. 13). They allow us to structure our thoughts, undertake logical trains of thought, and guide our behavior (Metzinger, 2013). The ability to perform second-order mental actions should hence form a key target for protection by the right to FoT.

​

Another issue relevant to developing the right to FoT is a move away from the conception that thought is only “in the head.” Clark and Chalmers (1998) make a persuasive case that the border between mind and world is not always constituted by the skin and skull. They propose that mind extends into the world. For them, the use of pen-and-paper to perform calculations or the re-arrangement of Scrabble tiles to make words, constitutes thinking. Clark and Chalmers argue that if “a part of the world functions as a process which, were it done in the head, we would have no hesitation in recognizing as part of the cognitive process, then that part of the world is (so we claim) part of the cognitive process” (p. 8). For example, in someone with dementia, a notebook could be functionally equivalent to memory.

​

In order for X to be classified as part of the extended mind, Clark and Chalmers (1998) propose three criteria: (1) X is a constant in the person's life and they will rarely take action without consulting it; (2) the information from X is directly available without difficulty; and, (3) upon retrieving information from X it is automatically endorsed. In 1998 Clark and Chalmers argued internet use was likely to fail to meet these criteria. Yet today, a case can be made that, when questions arise, much of the population will typically consult a search engine (e.g., Google), which they will always have access to without difficulty (via smartphones), which will lead them to trusted sources they will endorse. Internet searches have become thought.

​

The right to FoT should protect thought wherever it is found, not just in the head. This is important because if external thinking (e.g., internet searches, diaries, notebooks) is not deemed thought then it will only be protected by rights that can be permissibly violated under certain circumstances (e.g., the right to privacy). If it is deemed thought, then it will be protected by the absolute right of FoT. Any court that recognized the existence of external thought would be setting a highly impactful precedent.

​

Advances in Understanding Decision-Making

The extensive literature on factors that affect human decision making (Cialdini, 2013) includes the suggestion of two distinctly different systems that we possess for making decisions (Kahneman, 2011). System 1 (“rule-of-thumb” thought) is evolutionarily ancient, shared with animals, effortless and allows quick judgments to be made by using rules of thumb called heuristics. Its speed comes at the cost of accuracy. In contrast, system 2 (“rule of reason” thought) is a slow, consciously controlled process of reasoning, which is evolutionarily new, unique to humans, and linked to language. Its cost (slower decision making) can be offset by its increased accuracy. Attempts by governments and corporations to encourage citizens to use rule-of-thumb thinking, bypassing rule-of-reason thinking, could potentially violate the right to FoT.

​

Knowledge of what influences decision making is already being used by the state to influence behaviors. Governments “nudge” citizens into joining organ donor registers, increasing fruit and vegetable consumption, and increasing tax collection rates (Thaler and Sunstein, 2008; Marteau et al., 2011). Such “libertarian paternalism” is argued to still permit choice (and hence be ethical), even if it does bias individuals to make certain choices (Thaler and Sunstein, 2008). The commercial use of such knowledge to manipulate consumer's thoughts and behaviors is more ethically problematic (Ariely and Berns, 2010). Research into biases in human decision making have been utilized by the private sector to enhance the probability of consumers making purchases (Eyal, 2014). These techniques take advantage of foibles of the human mind, rather than engaging with rational faculties to encourage purchasing behavior. This arguably raises issues relating to FoT.

​

Social networking companies also make use of behavioral science insights. Their business models are based primarily on advertising income, motivating them to maximize the time people spend using their product. Sean Parker, the first president of Facebook, recently discussed the thinking that went into building this social network. He described it as being “All about how do we consume as much of your time and conscious attention as possible?” and doing this by “exploiting a vulnerability in human psychology…

 

The inventors, creators, it's me, it's Mark [Zuckerberg]… understood this consciously. And we did it anyway” (Pandey, 2017). Similarly, Guillaume Chaslot, a computer programmer who worked with YouTube engineers on its recommendation system, describes how “Watch time was the priority… Everything else was considered a distraction” (Lewis, 2018). Social networking sites achieve this by methods including variable reward reinforcement schedules (Ferster and Skinner, 1957), deployed in the context of providing information relevant to our fundamental need to belong (Baumeister and Leary, 1995) and desire for social status (Anderson et al., 2015). The imbalance in power resulting from the behavioral science knowledge of designers, whether they are creating social networks or casinos (Schüll, 2014), and the users of these products, raises the concern that users' FoT may be being violated.

​

In summary, our concept of what thought is and how the mind works have important implications for our conception of what the right to FoT should be. Before turning to a further consideration of these issues, there is the need to examine how the right to FoT is currently conceptualized.

​

The Right to Freedom of Thought

The right to FoT formally became international law in 1976 as part of Article 18 of the International Covenant on Civil and Political Rights (ICCPR). Today, it is a staple of most international human rights treaties. In Europe it is found in Article 9 of the European Convention on Human Rights (ECHR). In the United States, whilst elements of the Fourth and Fifth Amendment pertain to FoT, it is primarily the First Amendment that is understood to protect an individual's thought from Government interference (Richards, 2015). Although the First Amendment does not mention FoT explicitly, U.S. courts have noted that a concern for FoT was present at the time of its adoption and that it is closely associated with it (Kolber, 2016). U.S. courts have explicitly referred to a “First Amendment right to freedom of thought” (Doe v. City of Lafayette, Indiana, 2003) and the U.S. Supreme Court has stated that “at the heart of the First Amendment is the notion that an individual should be free to believe as he will” (Abood v. Detroit Board of Education, 1977).

​

Reflecting the fundamental importance of the right to FoT, both the ICCPR and ECHR give it the status of an absolute right. This means it cannot be interfered with in any circumstances. Article 4(2) of the ICCPR sets out the “fundamental character” of this right, which is “reflected in the fact that this provision cannot be derogated from, even in time of public emergency.” In the United States, the right to FoT is as close to an absolute right as any that exists in the Constitution (Blitz, 2010; Richards, 2015). However, whilst one view of the First Amendment is that it protects thought itself, another view is that it only protects thought when intertwined with expression (Kolber, 2016). Furthermore, U.S. courts have made statements that question whether the First Amendment can be seen as offering absolute protection (e.g., Branti v. Finkel, 1980). For example, Doe v. City of Lafayette, Indiana (2003), saw reference being made to a potential “exercise of determining what state interests might outweigh a person's right to think.”

​

The profound importance given to this right by the law contrasts with an almost complete lack of law clarifying what it actually entails. The United Nations Human Rights Committee (UNHRC) have stated that the scope of the right to FoT is “far-reaching and profound; it encompasses freedom of thoughts on all matters” (United Nations Human Rights Committee, 1993, italics added). Yet the right appears to be so profound as to deter any attempts at actually defining it. As Bublitz (2011) notes, “In its over 50 years of existence the ECHR has only decided a handful of cases regarding freedom of thought…[it] is an almost empty declaration. There are no definitions over its meaning, scope or possible violations” (p. 103). Similarly, the U.S. Supreme Court has never stated exactly what the right to FoT is (Blitz, 2010).

​

This lack of development appears to be because the mind has not traditionally been conceived as an entity vulnerable to external intrusions or interference and has hence not been viewed as in need of legal protection (Bublitz and Merkel, 2014; Richards, 2015). This view can be seen in 1942 when the U.S. Supreme Court declared that: “[F]reedom to think is absolute of its own nature, the most tyrannical government is powerless to control the inward workings of the mind.” When the Universal Declaration on Human Rights was drafted a few years later, delegates remarked: “It would be unnecessary to proclaim freedom of [the inner sphere] if it were never to be given an outward expression” as the inner is beyond any access (Hammer, as cited in Bublitz and Merkel, 2014).

​

A useful way to frame an exploration of the limited case law and scholarly writing on FoT is through Vermeulen's (2006) non-binding commentary on Article 9 of the ECHR. Vermeulen proposes the right to FoT has three key elements; a right not to reveal one's thoughts, a right not to have one's thoughts manipulated, and a right not to be penalized for one's thoughts. I will utilize this framework, exploring, for each element, descriptive aspects (what does law say about this right), normative aspects (why should this right exist), and contemporary challenges that will need to be taken into account when developing the right.

​

The Right not to Reveal One's Thoughts

Descriptive Aspects: Law as It Stands

In the United States the concept of mental privacy developed in the late nineteenth century from laws surrounding physical privacy. Warren and Brandeis (1890) argued for a right ‘to be let alone', which included protection for “thoughts, sentiments and emotions” (p. 205). Since this time U.S. courts have clarified that an individual's right to privacy (guaranteed by the Fourth Amendment) includes mental privacy. In Long Beach City Employees Assn. v. City of Long Beach (1986), California's Supreme Court stated that “[i]f there is a quintessential zone of human privacy it is the mind.” More generally, the U.S. Supreme Court has stated that “The Constitution protects the right… to be generally free from governmental intrusions into one's privacy and control of one's thoughts” (Stanley v. Georgia, 1969).

​

In Europe, Article 8 of the ECHR (“Right to respect for private and family life”) protects a space where one can be free from unwanted intrusions, whether that “be that the head or the home” (Marshall, 2009, as cited in Bublitz, 2013, p. 14). Mental privacy can also be understood as being protected by Article 9 of the ECHR (“Freedom of thought, conscience and religion”). Case law is sometimes unclear which of these Articles violations of mental privacy should be understood as involving (e.g., Folgerø Others v. Norway, 2007). Yet, it is important to be clear which ECHR Articles guarantee mental privacy. If given by the right to FoT, it will be an absolute right. If given by the right to privacy, it would have to be balanced against other concerns.

Normative Aspects: Why Should This Right Exist?

Mental privacy can be defined as our ability to determine for ourselves when, how, and to what extent information about our thoughts is communicated to others (cf. Westin, 1967). It reflects the idea there is a (contextually dependent) boundary between ourselves and others that must be respected (Richards, 2015). One argument for mental privacy is that it serves a protective function, defending us from distress at each other's less laudable thoughts and inclinations (Nagel, 1998). Another is that it facilitates intimacy, allowing us to choose to only bring certain individuals into our inner world, thereby acting as a marker of esteem or love (Nagel, 1998). However, the strongest argument for a right to mental privacy is that without it our mental autonomy would be diminished. “[A] lock on the door” noted Virginia Woolf (2001) “means the power to think for oneself” (p. 125). This is because exposure of our thoughts would effectively alter them by pressurizing us not to think certain things. As Cohen (2000) puts it, “Examination chills experimentation with the unorthodox, the unpopular, and the merely unfinished” (p. 1426).

​

The argument that mental privacy is necessary for mental autonomy rests on two claims. The first is that if we cannot keep our thoughts private then we will experience conformity pressures to think in a certain way. The second is that this pressure will cause self-censorship, impairing mental autonomy. What evidence is there for these claims?

​

We are clearly motivated to make our observable behaviors conform to social norms. Normative conformity (adjusting one's behavior in order to fit in with the majority) has been extensively documented and appears to exist because it serves three purposes (Cialdini and Goldstein, 2004). It generally enhances accuracy of decision making, enhances social approval leading to greater access to resources, and helps enhance, repair, and protect self-esteem (Cialdini and Goldstein, 2004). Conformity hence offers significant survival benefits.

​

As a result, our brains have evolved to motivate us to conform. The brain recognizes when we are deviating from social norms and reduces activation in areas associated with reward (e.g., the nucleus accumbens; Klucharev et al., 2009). Failing to conform is associated with increased activity in the amygdala, the brain's fear center (Berns et al., 2005), and in a number of other neural markers of stressful and aversive experiences (Yu and Sun, 2013). Thus, if our thoughts were made (or threatened to be made) public against our will, we would experience biologically rooted pressures to conform them to social norms.

​

Would such pressures lead our thoughts to conform? It is well-established that conformity pressures often result in people conforming their speech and behavior to group norms (Asch, 1952). Yet, behaviors are something we can reflect upon and potentially change. As we do not consciously choose our first-order thoughts, how could we make them conform to a standard?

​

Mental conformity could firstly occur by a conscious process of inhibiting thoughts as they arise. People use a range of techniques to control such thoughts, including punishment, worry, distraction, social control, and reappraisal (Wells and Davies, 1994). Yet mental conformity could also occur by non-conscious mechanisms. For example, Eriksen and Kuethe (1956) presented participants with a list of 15 words, and asked them to free associate other words related to these. Five of the words participants generated were followed by electric shocks. They were then asked to free associate with the same words again, but this time with no shock. It was found that participants showed a decrease in responding with the words that were punished by electric shock earlier. Some were not aware that the shocks were why they were doing this. Mental conformity can hence occur through an insidious, non-conscious process of self-censorship.

​

Evidence of self-censorship in a democracy as a result of violation of mental privacy can already be found if (as argued above) we accept that internet searches represent thought. A chilling effect on individual's internet search behavior was found to result from the June 2013 revelations of large-scale internet surveillance of individuals, as part of the information divulged by Edward Snowden. In this study, Marthews and Tucker (2017) collected data on internet search term volume before and after June 2013. They found that after this date there was a 10% drop in search terms that were assessed as potentially getting individuals in trouble with the government, as well as a significant drop in use of terms that could be personally embarrassing if revealed.

​

There is hence a fundamental need for what Richards (2015) has termed intellectual privacy, defined as a “zone of protection that guards our ability to make up our minds freely… the protection from surveillance or unwanted interference by others when we are engaged in the process of generating ideas” (p. 96). This should form an important pillar of the right to FoT.

​

Arguments against mental privacy are centered on benefits which pale by comparison to what would be lost if this right vanished. Richmond (2012) notes a potential benefit to making minds more transparent is the normalization of abhorrent first-order thoughts, reducing self-stigma and shame. Yet, this could equally well be done by other less intrusive means, such as public education (McCarthy-Jones, 2019).

 

Opponents of data privacy argue that the more information that marketers have, the better they can serve consumers and the less intrusive they need to be (Cohen, 2000). Both law and economics view one party having more knowledge than another as being inefficient as it hinders fully informed exchanges (Cohen, 2000). Privacy, by acting as a barrier to prediction (which is central to surveillance capitalism), becomes a barrier to efficiency. Yet, “a relentless focus on efficiency is the surest way to destroy liberty” (Rosen, 2016, p. 77). Furthermore, this case against mental privacy is an economic one, not a moral one. As Ryssdal and Garrova (2017) put it “The reason we don't have any privacy is because people can make money off of our not privacy.” It is not because it is in the interest of individuals to have no privacy. Indeed, without privacy we would, to a meaningful extent, not exist at all.

​

Contemporary Challenges That Should Shape the Right Going Forward

We have a natural ability to work out, to some degree, what others are thinking. This is termed “theory of mind” (ToM). In a typically developing child the basic form of ToM develops at around four-years-old (Wellman et al., 2001). The evolution of ToM, which only exists in its full blown form in humans, was presumably driven by the benefits of being able to predict another's behavior (Shultz and Dunbar, 2012).

 

As humans evolved the ability to work out others were thinking, this created selection pressures for the ability to conceal one's thoughts from another. This in turn put selection pressures on the ability to detect deception. It turns out that our ability to lie is significantly better than our ability to detect lying. Adults can only detect lies told by children at a rate slightly better than chance (Gongola et al., 2017). Even as children we have a powerful ability to encrypt our inner world. However, the development of brain- and behavior-reading techniques threaten to disturb the delicate balance between our evolved ability to know other's thoughts yet to also shield our inner world. But what exactly, as Ryberg (2017) has asked, “constitutes the difference between these two ways [natural and technological] of approaching other people's mental life”?

​

One answer is that the inappropriateness of brain- and behavior-reading stems from its supra-natural ability to infer a person's thoughts, steamrollering through evolved defenses, creating a significant power imbalance between the possessors of such technology and their “targets.” Power imbalances between the sender and recipient of a message has previously been ruled by the European Court of Human Rights to create an improper influence on thought (Larissis and Others v. Greece, 1998). The possessors of brain- and/or behavior-reading technology could hence be deemed to have “improper access” to the minds of other individuals. In addition, internet-use is so ubiquitous today that individuals may feel they have no choice but to surrender their data for behavior-reading, in exchange for services. This suggests a coercive element.

 

Another way to argue that brain- and behavior-reading is inappropriate, from a European perspective, is to start from the right to informational privacy given by EU Regulation 2016/679 (General Data Processing Regulation: GDPR). GDPR has no problems with one person naturally working out what another is thinking from their behavior (“data”) in everyday life. This would be legitimately processing someone's data as “a natural person in the course of a purely personal or household activity” [EU Regulation 2016/679, Art. 2, 2(c)]. However, if a company starts processing someone's data for the purposes of brain- or behavior-reading then they would need their explicit consent to do this. If they do not ask, or the person does not grant such consent, then such processing would be illegal. Of course, the GDPR has many permissible violations (e.g., national security).

​

If the right to FoT covered such invasions of mental privacy, no violations would be permissible.

 

This raises the question as to whether the right to mental privacy should be absolute. It is possible that society may wish permissible violations to exist. Warrants can be sought and granted to search individuals' homes. It is not immediately obvious that there should not be similar permissible violations for the mind. It is possible to imagine a judge granting a warrant for a mental search, which could take the form of an algorithmic analysis of all an individual's available data and/or the collection of neuroimaging data. The prosecution of hate crimes already often involve ransacking a person's life in order to determine what they were thinking (Corry, 2000).

​

Part of a justification for permissible violations could be the claim that the right to mental privacy needs to be balanced against other interests and rights. For example, Bublitz (2014) offers a thought experiment of a future in which outward signs of particular mental states (i.e., behavior-reading) could be observed to uncover individuals' thoughts in sensitive places such as banks, airports or even public parks. This would aim to deter the occurrence of such thoughts, leading to what Bublitz terms “zones of restricted freedom of thought” (p. 15). It is easy to see how this could slide into a dystopia, with people being punished for thoughts deemed dangerous. This highlights that most permissible violations of mental privacy would only make sense if people could justifiably be punished for such thoughts. But can they?

 

The Right not to be Penalized for One's Thoughts

Descriptive Aspects: The Law as It Stands

“[I]f there is any one proposition that commands general agreement among theorists and practitioners of the penal law,” Brudner (2009) has argued, “it is that judicial punishment ought not to be inflicted for private thoughts” (p. 108). Yet, in one sense it is common to punish thoughts. The famous maxim of criminal law states that an act does not make a defendant guilty without a guilty mind. In this sense, every crime with a mens rea requirement technically involves the person being punished for thoughts. Similarly, in the case of hate crime, Corry (2000) has argued that “[b]ecause these laws seek to punish certain unacceptable prejudiced thoughts, more accurate terms for these laws are ‘thought crimes”' (p. 469).

 

Yet, it would not be correct to say people are being punished for “mere thought” in such cases. As was noted in U.S. v. Gamache (1998), “[p]roof of intent naturally means proving state of mind, but that does not mean that one is punishing ‘mere thought' any more than that the requirement of proving mens rea in most crimes means that one is solely punishing ‘mere thought.”' Some substantial step has to have been taken toward a criminal act, in addition to the thought, comprising “objective acts [which]…mark the defendant's conduct as criminal in nature,” meaning that a “fragment of the crime” is in progress (U.S. v. Price, 1998). U.S. courts have stated that the fear “that thoughts alone may encourage action—is not enough to curb protected thinking” (Doe v. City of Lafayette, Indiana, 2003). In the absence of any substantial steps toward a criminal act, law in the United States recognizes “the fundamental constitutional principle that a person's thoughts are his own—however distasteful they may be to the state or to the populace” (Steffan v. Perry, 1994). Mere thought is hence protected.

 

Yet, in prosecutions for attempted crimes it is possible to conceive that the person is being punished for a thought; namely, an intention (Austin, 1885). The substantial steps requirement only means that stronger evidence has been gained of the strength of the intention, it does not mean that something other than the intention is being punished (Morris, 1965). Yet, it can be argued that what the person is being punished for is conduct accompanied by a certain state of mind, which makes it illegal conduct.

​

A series of cases involving Doe v. City of Lafayette, Indiana (2001, 2003, 2004) are central to present concerns because they gave rise to “a set of facts that tests just how far the judiciary is willing to stretch and extend the unenumerated constitutional right of freedom of thought when the thoughts in question relate to committing one of the most heinous of crimes—child molestation” (Calvert, 2005, p. 138). Doe, a convicted sex offender, was driving home from work when he began to have sexual thoughts about children. He drove to a city park, watched children, had thoughts about having sexual contact with them, and then left without having any such contact. An anonymous source (potentially from his Sexual Addicts Anonymous group; Calvert, 2005) later reported Doe's thoughts to officials. The City decided, although Doe was no longer serving a sentence or on probation, to ban him from entering any City parks.

​

Doe challenged the ban, which a district judge upheld (Doe v. City of Lafayette, Indiana, 2001). Doe then appealed this decision to a three judge panel of the U.S. Seventh Circuit Court of Appeals (Doe v. City of Lafayette, Indiana, 2003). The Court concluded that Doe's actions did not “reach a level of criminal culpability necessary to justify punishment” and that his simple presence in the park, thinking sexual thoughts about children, did not qualify as a “substantial step” toward an attempted sex offense. They hence decided that the ban violated “Doe's First Amendment right to freedom of thought” by punishing him for “pure thought,” and overturned it. Eleven judges of the Seventh Circuit Court of Appeals then reheard the case (Doe v. City of Lafayette, Indiana, 2004). This time the court upheld the ban, reasoning Doe was banned from parks because of his actions, not because of his thoughts. Whether or not thought can be punished hence appears to be somewhat open to interpretation.

Normative Aspects: Why Should This Right Exist?

In the nineteenth century, Sir William Blackstone (1832) stated that “as no temporal tribunal can search the heart, or fathom the intentions of the mind, otherwise than as they are demonstrated by outward actions, it therefore cannot punish for what it cannot know” (p. 14). As technological advances make the reliable assessment of inner states increasingly feasible, arguments for not punishing thought now need to be solidly based on principle.

​

The most influential, principle-based argument against punishing thought comes from the following syllogism. Major premise: We can only punish that which violates important interests of others in which they have a right. Minor premise: Thoughts cannot violate important interests of others in which they have a right. Conclusion: We cannot not punish thoughts. The major premise is termed the “harm principle.” This stems from John Stuart Mill's argument that the “only purpose for which power can be rightfully exercised over any member of a civilized community, against his will, is to prevent harm to others” (Mill, 1859/1985, p. 68).

​

The ex-post version of the harm principle states that coercion is justified only if it actually prevents or reduces harm, but in its ex-ante version coercion is justified if it prevents or reduces risk (Holtug, 2002). The difference is important. The question of whether a thought actually causes harm to others is very different to whether a thought increases the risk of harm to others. If reckless driving can be prosecuted as it increases risk that the law seeks to diminish (Feinberg, 1987), why could we not prosecute reckless thinking? It turns out that thought's importance to society, rather than its innocuousness, saves it.

​

The contention that thought cannot risk others is something of a strawman. First, it would be contradictory to say that we need FoT to be able to control our behavior, but that our thoughts can't risk anyone. Second, a core tenet of contemporary evidence-based psychology is that our thoughts can harm us and others through influencing our behaviors, emotions and physiology (Beck and Haigh, 2014). Third, if overt speech can incite others to act and thereby cause harm, then it logically follows that our covert, inner speech can incite ourselves to act and thereby risk others. The Russian psychologist Lev Vygotsky (1978) was clear that a central purpose of our inner speech is to allow children to self-regulate their behavior and thoughts. As he put it, the use of verbal mediation means that humans are able to “control their behavior from the outside” (p. 40). To completely decontextualize thought from action is to fundamentally misunderstand its purpose.

Better questions are whether different thoughts are associated with different levels of risk, which have most risk, and how great this risk is. It can be argued that risk of harm from thoughts is typically low.

 

The psychological research literature is replete with examples of people's beliefs being poor predictors of their behavior (e.g., LaPiere, 1934). Whilst some have expressed extreme pessimism at the ability of attitudes to predict behaviors (Wicker, 1969), the research literature on this topic is complex (Kruglanski et al., 2015). Nevertheless, it is generally the case the attitudes predict behavior to a lesser degree than we might expect.

Yet some categories of thoughts clearly pose more risk of harm than others. The theory of planned behavior argues that behavior can be best predicted from intention (Ajzen, 1991). Indeed, as Mendlow (2018) observes

 

“There would be little point to forming intentions if intentions didn't generally increase the likelihood of actions.” (p. 2017). Not only can intentions create risk of harm but, as Mendlow observes, in the case of an intention to kill formed after extensive reflection and deliberation they can arguably be associated with a comparable risk of harm to actions that we already criminalize on account of their dangerousness, such as driving recklessly or possessing volatile explosives. Duff (2007) also suggests a potentially harmful category of thoughts are those for which the completion requires overt action, with intention formation being an example of this.

​

Another type of thought that increase risk of harm are second-order thoughts that endorse and encourage unwanted, unacceptable first-order thoughts. Someone who fails to find unwanted, unacceptable first-order thoughts repugnant, who does not try to suppress or avoid them, who does not try to avoid situations that trigger them, or is aroused or acts on them, may lead us to have cause for concern (Veale et al., 2009). Someone who felt this way about intrusive sexual thoughts they had about children is a potential sexual offender (Veale et al., 2009). This can be contrasted with Doe's claimed second-order thoughts in Doe v. City of Lafayette, Indiana (2004), which rejected his first-order sexual thoughts (“I was not planning to act on my thoughts. I recognized that these were just unhealthy thoughts and I realized I needed to leave the park, which is what I did”).

​

Even though thoughts can increase risk of harm, to punish them it would still need to be shown that the thinker merited condemnation or blame, i.e., was culpable for them (Mendlow, 2018). In the case of an intention to kill, reached after extensive deliberation (second-order thought), condemnation and blame are clearly deserved (Mendlow, 2018). Yet, we are not culpable for all our potentially harmful thoughts. Unwanted, intrusive and abhorrent first-order thoughts are common (Purdon and Clark, 1993; Byers et al., 1998) but one should not be deemed culpable for them (McCarthy-Jones, 2019).

​

Having demonstrated that some thoughts can be both dangerous and culpable, a remaining objection to punishing thought stems from the gap between intent and action. Punishing at the stage of mere intent diminishes the autonomy of the person by not allowing them to self-govern and restrain themselves from performing the act. Waiting for significant preparatory steps to be taken gives people the maximum autonomy to change their mind, before the law steps in coercively (Morris, 1965). However, Mendlow (2018) notes that such arguments assume that someone “hasn't already committed a punishable wrong simply by preparing or intending.” (p. 2018, italics in the original). He also notes that “To intervene on the expectation that he'll follow through on his criminal choice is arguably to show respect for his capacity for self-government.” (p. 2365).

 

However, a practical reason not to punish intentions comes from Morris (1965). He notes that if people were to be punished simply for their intentions then, once their intention had formed, they would have no incentive not to perform the act. Not punishing mere intentions rewards self-restraint.

All this brings us to the point where we now need to explain, as Mendlow (2018) puts it, “why it's intrinsically unjust to punish mental states that are provable, dangerous, and culpably wrongful: mental states that bear the chief hallmarks of paradigmatic punishable actions (p. 2360). Mendlow's answer is complicated, and the reader is referred to his excellent treatment of this question. His basic idea is that punishment for mere mental states is intrinsically unjust because it represents, indirectly, mind control, and “it's because the state mustn't control thoughts that the state mustn't punish them” (p. 2018).

​

Another argument against punishing thought stems from a consideration of what is required for related rights, such as the right to freedom of speech, to be limited. Article 19 of the ICCPR states that limitations may be placed on freedom of expression in certain cases and if one can “demonstrate in specific and individualized fashion the precise nature of the threat, and the necessity and proportionality of the specific action taken, in particular by establishing a direct and immediate connection between the expression and the threat” (United Nations Human Rights Committee, 2011). In the United States, the Brandenburg test stipulates that speech can only be limited if it is “directed to inciting or producing imminent lawless action and is likely to incite or produce such action” (Brandenburg v. Ohio, 1969). It is likely to be hard to show an immediate/imminent connection between a thought and a threat.

​

Furthermore, when we consider how strong the substantial action test is for establishing “attempt,” in the case of observable behaviors in the United States, it is hard to see how mere thought could be deemed attempt. For example, as Duff (1998) notes, in United States v. Still, the defendant was found wearing a disguise whilst parked in a van 200 feet from a bank. After his arrest, the defendant said to the police “You did a good job. You caught me five minutes before I was going to rob a bank. That's what I was putting the wig on for.” And still the defendant was not convicted of attempt. If preparations this clear cannot be deemed attempt, it is hard to see how the thoughts of a defendant, on their own, could ever be deemed attempt and punished.

Challenges That Should Shape the Right

In time, greater access to the thoughts of individuals, through brain- and behavior-reading, will raise the question of whether, once we can know such thoughts with an appropriate degree of certainty, we should punish them. If, despite the above arguments against punishing thoughts, the public and their elected representatives were to support punishment for thought, why might his be? One answer is that this would be being done “to calm the anxiety of others for their physical security or psychic integrity, or simply to achieve dominance for their moral opinions” (Brudner, 2009, p. 108).

​

Even without new legislation, judges' views on the punishment of thought may be influenced by a risk-adverse public clamoring for such punishment. As the former Justice of the Supreme Court of the United Kingdom, Lord Sumption (2019), has noted, “Judges don't decide cases in accordance with the state of public opinion but it is their duty to take account of the values of the society which they serve. Risk aversion has become one of the most powerful of those values and is a growing influence in the development of the law.”