In 1984, the philosopher Aaron Sloman invited scholars to describe ‘the space of possible minds’. Sloman’s phrase alludes to the fact that human minds, in all their variety, are not the only sorts of minds. There are, for example, the minds of other animals, such as chimpanzees, crows and octopuses. But the space of possibilities must also include the minds of life-forms that have evolved elsewhere in the Universe, minds that could be very different from any product of terrestrial biology. The map of possibilities includes such theoretical creatures even if we are alone in the Cosmos, just as it also includes life-forms that could have evolved on Earth under different conditions.
We must also consider the possibility of artificial intelligence (AI). Let’s say that intelligence ‘measures an agent’s general ability to achieve goals in a wide range of environments’, following the definition adopted by the computer scientists Shane Legg and Marcus Hutter. By this definition, no artefact exists today that has anything approaching human-level intelligence. While there are computer programs that can out-perform humans in highly demanding yet specialised intellectual domains, such as playing the game of Go, no computer or robot today can match the generality of human intelligence.
But it is artefacts possessing general intelligence – whether rat-level, human-level or beyond – that we are most interested in, because they are candidates for membership of the space of possible minds. Indeed, because the potential for variation in such artefacts far outstrips the potential for variation in naturally evolved intelligence, the non-natural variants might occupy the majority of that space. Some of these artefacts are likely to be very strange, examples of what we might call ‘conscious exotica’.
In what follows I attempt to meet Sloman’s challenge by describing the structure of the space of possible minds, in two dimensions: the capacity for consciousness and the human-likeness of behaviour. Implicit in this mapping seems to be the possibility of forms of consciousness so alien that we would not recognise them. Yet I am also concerned, following Ludwig Wittgenstein, to reject the dualistic idea that there is an impenetrable realm of subjective experience that forms a distinct portion of reality. I prefer the notion that ‘nothing is hidden’, metaphysically speaking. The difficulty here is that accepting the possibility of radically inscrutable consciousness seemingly readmits the dualistic proposition that consciousness is not, so to speak, ‘open to view’, but inherently private. I try to show how we might avoid that troubling outcome.
Thomas Nagel’s celebrated treatment of the (modestly) exotic subjectivity of a bat is a good place to start. Nagel wonders what it is like to be a bat, and laments that ‘if I try to imagine this, I am restricted to the resources of my own mind, and those resources are inadequate to the task’. A corollary of Nagel’s position is that certain kinds of facts – namely facts that are tied to a very different subjective point of view – are inaccessible to our human minds. This supports the dualist’s claim that no account of reality could be complete if it comprised only objective facts and omitted the subjective. Yet I think the dualistic urge to cleave reality in this way is to be resisted. So, if we accept Nagel’s reasoning, conscious exotica present a challenge.
But bats are not the real problem, as I see it. The moderately exotic inner lives of non-human animals present a challenge to Nagel only because he accords ontological status to an everyday indexical distinction. I cannot be both here and there. But this platitude does not entail the existence of facts that are irreducibly tied to a particular position in space. Similarly, I cannot be both a human and a bat. But this does not entail the existence of phenomenological facts that are irreducibly tied to a particular subjective point of view. We should not be fooled by the presence of the word ‘know’ into seeing the sentence ‘as a human, I cannot know what it’s like to be a bat’ as expressing anything more philosophically puzzling than the sentence ‘I am a human, not a bat.’ We can always speculate about what it might be like to be a bat, using our imaginations to extend our own experience (as Nagel does). In doing so, we might remark on the limitations of the exercise. The mistake is to conclude, with Nagel, that there must be facts of the matter here, certain subjective ‘truths’, that elude our powers of collective investigation.
to explore the space of possible minds is to entertain the possibility of beings far more exotic than any terrestrial species
In this, I take my cue from the later Wittgenstein of The Philosophical Investigations (1953). The principle that underlies Wittgenstein’s rejection of private language – a language with words for sensations that only one person in the world could understand – is that we can talk only about what lies before us, what is public, what is open to collective view. As for anything else, well, ‘a nothing would serve as well as a something about which nothing can be said’. A word that referred to a private, inner sensation would have no useful function in our language. Of course, things can be hidden in a practical sense, like a ball beneath a magician’s cup, or a star that is outside our light cone. But nothing is beyond reach metaphysically speaking. When it comes to the inner lives of others, there is always more to be revealed – by interacting with them, by observing them, by studying how they work – but it makes no sense to speak as if there were something over and above what can ever be revealed.
Following this train of thought, we should not impute unknowable subjectivity to other people (however strange), to bats or to octopuses, nor indeed to extra-terrestrials or to artificial intelligences. But here is the real problem, namely radically exotic forms of consciousness. Nagel reasonably assumes that ‘we all believe bats have experience’; we might not know what it is like to be a bat, yet we presume it is like something. But to explore the space of possible minds is to entertain the possibility of beings far more exotic than any terrestrial species. Could the space of possible minds include beings so inscrutable that we could not tell whether they had conscious experiences at all? To deny this possibility smacks of biocentrism. Yet to accept it is to flirt once more with the dualistic thought that there is a hidden order of subjective facts. In contrast to the question of what it is like to be an X, surely (we are tempted to say) there is a fact of the matter when it comes to the question of whether it is like anything at all to be an X. Either a being has conscious experience or it does not, regardless of whether we can tell.
Consider the following thought experiment. Suppose I turn up to the lab one morning to discover that a white box has been delivered containing an immensely complex dynamical system whose workings are entirely open to view. Perhaps it is the gift of a visiting extraterrestrial, or the unwanted product of some rival AI lab that has let its evolutionary algorithms run amok and is unsure what to do with the results. Suppose I have to decide whether or not to destroy the box. How can I know whether that would be a morally acceptable action? Is there any method or procedure by means of which I could determine whether or not consciousness was, in some sense, present in the box?
One way to meet this challenge would be to devise an objective measure of consciousness, a mathematical function that, given any physical description, returns a number that quantifies the consciousness of that system. The neuroscientist Giulio Tononi has purported to supply just such a measure, named Φ, within the rubric of so-called ‘integrated information theory’. Here, Φ describes the extent to which a system is, in a specific information-theoretic sense, more than the sum of its parts. For Tononi, consciousness is Φ in much the same sense that water is H2O. So, integrated information theory claims to supply both necessary and sufficient conditions for the presence of consciousness in any given dynamical system.
The chief difficulty with this approach is that it divorces consciousness from behaviour. A completely self-contained system can have high Φ despite having no interactions with anything outside itself. Yet our everyday concept of consciousness is inherently bound up with behaviour. If you remark to me that someone was or was not aware of something (an oncoming car, say, or a friend passing in the corridor) it gives me certain expectations about their behaviour (they will or won’t brake, they will or won’t say hello). I might make similar remarks to you about what I was aware of in order to account for my own behaviour. ‘I can hardly tell the difference between those two colours’; ‘I’m trying to work out that sum in my head, but it’s too hard’; ‘I’ve just remembered what she said’; ‘It doesn’t hurt as much now’ – all these sentences help to explain my behaviour to fellow speakers of my language and play a role in our everyday social activity. They help us keep each other informed about what we have done in the past, are doing right now or are likely to do in the future.
It’s only when we do philosophy that we start to speak of consciousness, experience and sensation in terms of private subjectivity. This is the path to the hard problem/easy problem distinction set out by David Chalmers, to a metaphysically weighty division between inner and outer – in short, to a form of dualism in which subjective experience is an ontologically distinct feature of reality. Wittgenstein provides an antidote to this way of thinking in his remarks on private language, whose centrepiece is an argument to the effect that, insofar as we can talk about our experiences, they must have an outward, public manifestation. For Wittgenstein, ‘only of a living human being and what resembles (behaves like) a living human being can one say: it has sensations, it sees … it is conscious or unconscious’.
Through Wittgenstein, we arrive at the following precept: only against a backdrop of purposeful behaviour do we speak of consciousness. By these lights, in order to establish the presence of consciousness, it would not be sufficient to discover that a system, such as the white box in our thought experiment, had high Φ. We would need to discern purpose in its behaviour. For this to happen, we would have to see the system as embedded in an environment. We would need to see the environment as acting on the system, and the system as acting on the environment for its own ends. If the ‘system’ in question was an animal, then we already inhabit the same, familiar environment, notwithstanding that the environment affords different things to different creatures. But to discern purposeful behaviour in an unfamiliar system (or creature or being), we might need to engineer an encounter with it.
Even in familiar instances, this business of engineering an encounter can be tricky. For example, in 2006 the neuroscientist Adrian Owen and his colleagues managed to establish a simple form of communication with vegetative-state patients using an fMRI scanner. The patients were asked to imagine two different scenarios that are known to elicit distinct fMRI signatures in healthy individuals: walking through a house and playing tennis. A subset of vegetative-state patients generated appropriate fMRI signatures in response to the relevant verbal instruction, indicating that they could understand the instruction, had formed the intention to respond to it, and were able to exercise their imagination. This must count as ‘engineering an encounter’ with the patient, especially when their behaviour is interpreted against the backdrop of the many years of normal activity the patient displayed when healthy.
we don’t weigh up the evidence to conclude that our friends are probably conscious creatures. We simply see them that way, and treat them accordingly
Once we have discerned purposeful behaviour in our object of study, we can begin to observe and (hopefully) to interact with it. As a result of these observations and interactions, we might decide that consciousness is present. Or, to put things differently, we might adopt the sort of attitude towards it that we normally reserve for fellow conscious creatures.
The difference between these two forms of expression is worth dwelling on. Implicit in the first formulation is the assumption that there is a fact of the matter. Either consciousness is present in the object before us or it is not, and the truth can be revealed by a scientific sort of investigation that combines the empirical and the rational. The second formulation owes its wording to Wittgenstein. Musing on the skeptical thought that a friend could be a mere automaton – a phenomenological zombie, as we might say today – Wittgenstein notes that he is not of the opinion that his friend has a soul. Rather, ‘my attitude towards him is an attitude towards a soul’. (For ‘has a soul’ we can read something like ‘is conscious and capable of joy and suffering’.) The point here is that, in everyday life, we do not weigh up the evidence and conclude, on balance, that our friends and loved ones are probably conscious creatures like ourselves. The matter runs far deeper than that. We simply see them that way, and treat them accordingly. Doubt plays no part in our attitude towards them.
How do these Wittgensteinian sensibilities play out in the case
