Functionalist theories of mind

Functionalism is a, if not the, leading approach to the metaphysics of mind in contemporary philosophy.Footnote ⁹ Functionalism is often introduced with obvious functional kinds: barriers, vehicles, traps, pumps, locks, calculators, leaders, law enforcement officers, and so on. These are defined not in terms of the stuff they are made of (silver, steel, plastic, silicon, grey matter, human bodies, souls, or some combination thereof), but in terms of the functional role they fulfil in a system. The kind of stuff that instances of them are made of only matters insofar as that kind of stuff is required to fulfil a given role in a system.

Let’s walk through, at a rather high level, how functionalists think of particular mental states before mentioning their view of minds and persons. Take Socrates’s belief that it is raining. Beliefs are a functional kind, a kind whose essential characteristics involve: specific ways of interacting with evidential inputs (e.g., perceiving rain, hearing weather reports, feeling raindrops); specific ways of interacting with other mental states (e.g., the desire not to get wet); specific ways of contributing to the formation of intentions (e.g., intending to take an umbrella or remain indoors); and a tendency to produce corresponding behaviours under certain conditions (e.g., retrieving an umbrella or choosing not to go outside). These ‘specific ways of interacting’ are to be specified largely in terms of counterfactual dependence relations. So a belief that it is raining is a type of mental state that is to be identified with that type of functional role, while Socrates’s token belief is the realisation of that functional role in Socrates. The realisation of mental states in Socrates happens in virtue of Socrates instantiating a lower-level or ‘base’ property (e.g., having a certain neuronal arrangement) that has the functional role of the mental state in question.Footnote ¹⁰

A distinctive characteristic of functionalism is the multiple realisability of mental states: the kind of substance that makes up a system (biological, artificial, immaterial souls) does not matter so long as the system can instantiate the lower-level properties that realise mental states. Accordingly, functionalism implies that mental states can be not only realised in Socrates, but also re-realised in Socrates. Take a case of redundant memory storage. All that is needed for a re-realisation of a memory to occur in Socrates is for Socrates to twice instantiate the lower-level property that realised that memory to begin with, or for Socrates to instantiate a different lower-level property that realises that memory’s characteristic functional profile. Hence multiple realisability can occur within single individuals as well as across distinct individuals.

So functionalism treats mental state types as abstract role types, and it treats concrete individuals with minds as systems that instantiate token mental states in virtue of instantiating lower-level realisers, for example certain neuronal arrangements. So any system that instantiates mental states has a mind in at least a minimal sense. Like mental state types, mind-types are individuated by functional organisation: structured networks of functional roles. A collection of token states m₁…m_n (e.g., beliefs, desires, intentions, etc.) belongs to one mind just when m₁…m_n stand in the right pattern of mutual dependence (counterfactual/control) relations that integrate belief, desire, intention, etc. A single system can instantiate many mental state tokens without those tokens all being part of one mind. Functionalists can say, for example, that this is what separates the beliefs of Jekyll from those of Hyde. Jekyll and Hyde may be distinct minds instantiated in a single system (e.g., one brain) owing to a lack of functional integration among their mental states.

What is the relation between systems and substances (i.e., the bearers of properties)? The individual substance(s) that yield a mind (i.e., a mental state instantiating system) can in principle be a collection of distinct substances, a single mereologically composite substance, or a single mereologically non-composite (simple) substance. All that matters is that the substance(s) in question instantiate the lower-level properties that realise the relevant functional roles.Footnote ¹¹

Minds may fail to be persons. Newborn lambs realise mental states given that they see, hear, and respond to things in their environment. But they fall short of the robust kind of personhood that is characteristic of mature human beings and divine persons. While characterisations of this sort of personhood can differ, such persons are widely taken to be minds that are conscious, self-conscious, and have rational capacities.Footnote ¹² The relevant rational capacities are generally treated as a matter of having a constellation of functionally characterisable abilities involving reasons.Footnote ¹³ As such persons are conscious, self-conscious, and rational beings, they are not abstract types that can be multiply realised. But mind-types and person-types are abstract kinds and these are multiply realisable. Were God to have cloned Adam rather than making Eve, he would have brought about two realisations of the person-type of Adam. There would then have been two token persons, but only one person-type.

Not all who are sympathetic to the functionalist account of mental states, minds, and persons embrace it without qualification. For example, Chalmers (Reference Chalmers1996, 247–249) has argued for non-reductive functionalism on which only non-phenomenal mental states can be identified with functional states, while phenomenal/conscious mental states are to be seen as systematically correlated with functional mental states. Alternatively, some might wish to maintain that the functionalist framework that fits best with traditional monotheism is one that adds the constraint that only immaterial substances (souls) are capable of realising the functional organisations sufficient for genuine mental states. This would be an anti-physicalist functionalism. Importantly, the view of the trinity to emerge below is consistent with these deviations from standard functionalism, and for expressive convenience we will limit our attention to standard functionalism.

From functionalism to the second person of the trinity

The idea that God¹ could cause or bring about a distinct co-equal person is not new to philosophical theology. Following Augustine (Reference Matthews2012) and Richard of St. Victor (1958), Swinburne (Reference Swinburne2018, 429) takes this to be a straightforward implication of God¹ having maximal power. But we do not want to take this for granted without discussion. For we want to establish the claim that a being like God¹ would have what it takes to bring about the re-realisation of his own person-type in his own system.

According to traditional monotheistic views, God¹ has maximal knowledge. This includes both propositional knowledge (knowledge of truths) as well as practical knowledge (knowledge of how to do things). Now, God’s propositional knowledge extends to God¹ himself, that is, God¹ knows all the truths it is broadly logically possible for him to know about himself.Footnote ¹⁴ Now, standard views of propositional knowledge entail that persons have propositional knowledge because they have certain representational mental states (e.g., beliefs).Footnote ¹⁵ Assuming the same is true of God¹’s propositional knowledge, God¹’s propositional self-knowledge is constituted by a collection of self-representational mental states.

It is implausible to maintain that the collection of God¹’s self-representational mental states themselves realise the functional role persons.Footnote ¹⁶ But, according to functionalism, God¹’s propositional self-knowledge and practical knowledge give him an opportunity. For God¹ is a person who is himself realised in a particular system (presumably an immaterial substance). And God¹’s knowledge will not only include facts about his mental states, it will also include knowledge of facts about how his mental states are realised in his system as well as knowledge how to bring about realisations of minds like his own within systems like his own. To assume otherwise would be to assume that God¹’s propositional and practical knowledge is not maximal.

Knowing all the facts about how a mind is related to its system as well as knowing how to bring about a realisation of a mind in a system is not generally enough to ensure that one has the ability or power to do so.Footnote ¹⁷ For example, when we are considering the possibility of realising two minds within a single existing system, the issue becomes one of the capacity of the system in question. Some might worry, then, that God¹’s system is already ‘maxed out’ by God¹ himself and thus is not capable of enabling a re-realisation of God¹’s person-type. This is a worry that has bite with finite systems: for if the capacities of your system have an upper bound, it may not be possible for it to enable a re-realisation of your person-type.

But this worry gains less traction in the case of infinite systems, that is, systems that have the capacity to realise infinitely many states and dependence relations among those states. That is, assume the number of states and dependency relations among the states capable of being realised in God’s system is some order of infinity ℵ_n. Any system capable of realising ℵ_n-many states and dependency relations has the capacity to host 2 ✕ ℵ_n such states and dependency relations. This is an entailment of the mathematically trivial result that for any natural number к and any infinite quantity ℵ_n: к ✕ ℵ_n = ℵ_n.Footnote ¹⁸ Since the natural numbers are countably infinite, you can select any arbitrarily large finite number к within the natural numbers and the result of multiplying it by ℵ_n will remain ℵ_n. Thus, for any natural number к, if God¹ is realised in an infinite system, then God¹’s system has the capacity to host к-many more realisations of God¹’s person-type.Footnote ¹⁹ We will assume in what follows that God¹’s maximal knowledge and power require that God¹’s person-type be realised in an infinite system in the sense noted above. Within the space of monotheistic views, it is not unusual to maintain that God is infinite in various senses. What is unusual is the way it is being recast in functionalist terms: a mind with infinite capacity requires the system that instantiates it to have infinite capacity.

Take the following collection of claims:

The Maximality Package appears to entail the following:

If God¹ has this power, then the result of God¹ exercising such power would be a token-distinct person according to functionalism. Again, let’s refer to this prospective second token instance of God¹’s person-type as God².

For the Maximality Package to fail to entail PPSR, it would have to be the case that God¹’s power, knowledge, or system could be limited in some way that either (i) prevented any additional realisation of his person-type, or (ii) prevented a re-realisation of his person-type that is maximally accurate. Standard views about God¹’s extremely high degree of knowledge and power will not support blocking the entailment at (i). But blocking the entailment at (ii) requires an unintended reading of ‘maximal accuracy’. So we’ll now say a couple of words about that.

To claim that God² is a maximally accurate re-realisation of God¹ is not to claim that they share all their intrinsic features. For God¹’s maximal power and self-knowledge can only ensure that there are no avoidable differences between God¹ and God². But some differences cannot be avoided. For example, if God² is to have maximal knowledge he should believe no falsehoods. Thus, he should not, for example, believe that he is numerically identical to God¹ even though God¹ will believe that he is identical to God¹. In general, God² should not believe falsehoods about who he is, who God¹ is, and what relations they stand in to each other. All of this logically requires that God¹ and God² have some distinct mental states.Footnote ²⁰ This is all consistent with the intended reading of ‘maximal accuracy’.

Now, if God¹ re-realises his own person-type in his own system in a way that’s maximally accurate, then God²’s powers cannot fail to be as expansive as his own. For functionalism seems to imply that God¹ has the powers that he has because of the kind of person that he is and because of the kind of system his mind is realised in. Since God² is a maximally accurate re-realisation of God¹ in the same system, it follows that God² and God¹ should share the same powers.

For that same reason, God¹ and God² will be alike in their power to freely act and this will bring about further unavoidable differences between God¹ and God². For if God² is a maximally accurate re-realisation of God¹, and God¹’s power to act freely is a product of God¹ being the kind of person that he is, then God² must enjoy the same degree of freedom that God¹ does. Thus, when it comes to purely discretionary actions (free actions that are not required by God¹’s nature) God¹ and God² might make different choices or have (or develop) different preferences. So in addition to having some distinct beliefs, God¹ and God² may well develop distinct preference states. Again, all of this is consistent with the intended reading of ‘maximal accuracy’.

Let’s quickly summarise where we’re at. This section has argued that if functionalism is true, then all persons exist in virtue of systems having lower-level properties that realise the functional profile of a person-type. The same is true of God¹. And since he is a person of maximal power and maximal (propositional and practical) knowledge in an infinite system, God¹ has the capacity to bring about a second, maximally accurate realisation of his own person-type in his own system.

Collective person realism

Can a collection of distinct persons interact in ways that yield a further distinct person? For present purposes, answering this question in perfect generality is irrelevant. What is relevant is what follows from functionalism. That is, what we are concerned with is whether the following conditional is true:

The qualification ‘individual’ is meant to highlight an unintended reading of CPR on which performative (/group, /collective) persons are said to exist in virtue of the activities of individual persons. Corporations, administrations, school boards, and the like are examples of performative persons, each of which exists in virtue of the activities of their individual members.Footnote ²¹

But performative persons cannot easily rescue trinitarians from their trilemma for two reasons. First, it’s not clear that performative persons, in their familiar manifestations, have all the essential properties of individual persons. Corporations, administrations, school boards, and the like are not plausibly regarded as conscious and self-conscious. So they are not ‘persons’ in the way required for trinitarianism. Second, for most trinitarians, all members of the trinity are supposed to be persons in the same sense of ‘person’. So even though one could argue for a non-standard version of trinitarianism on which the third person of the trinity is merely a non-conscious performative person, this would be attractive to very few trinitarians.

Now, CPR is true. Functionalists and anti-functionalists have noted this. To appreciate this it helps to consider Ned Block’s (Reference Block1978) ‘China Brain’ thought experiment, which involves a case like this:

This kind of case was issued as an objection to functionalism: a functional organisation involving lots of individual people with hand-held radios seems incapable of yielding a mind that has conscious experiences, for example experiences of pain, pleasure, colour, hope, desire, and the like.

But in defence of functionalism, Chalmers (Reference Chalmers1996) has argued that cases like the Socially Realised Brain are highly under-described, representing only a caricature of the functionalist’s position. Indeed, when cases like the Socially Realised Brain are resolved into detail sufficient to accurately represent the functionalist view, the intuition that a nation of people cannot realise the same kind of mind that a brain does loses its force. He writes:

This defence of functionalism along with Block’s objection ratifies, rather than resists, CPR. For functionalism entails that the Chinese nation or homunculi can bring about a realisation of the function of an individual person. Readers sceptical of functionalism owing to the entailment of CPR and the possibility of the Socially Realised Brain are encouraged to remember that the purpose of this paper is not to convince you of functionalism; rather, the point is to convince you of what follows for trinitarianism if functionalism is consistent and conceivable. At any rate, readers uncomfortable with standard functionalism at this point might be more comfortable thinking about what follows in terms of some anti-physicalist version of functionalism on which only immaterial substances can realise minds.

Towards the third person of the trinity

CPR is good news for the trinitarian. For CPR entails that: if (i) God¹ brought about a maximally accurate re-realisation of his person-type exactly once (yielding God²) and if (ii) God¹ and God² interacted in ‘the right kind’ of ways, then that could itself bring about a further re-realisation of God¹. Let’s call this potential third person God³.

To see that CPR entails this, recall the Socially Realised Brain. In that case we have a group of individual human persons that engage in coordinated activities in ways that ensure the existence of a further person whose internal psychological and phenomenal states are just like that of some arbitrary human person. Let us alter the case somewhat to make it more applicable to trinitarian concerns:

The thing to keep in mind is that Socially Realised Adam is just a special instance of Socially Realised Brain. So if Socially Realised Brain is entailed by functionalism, so is Socially Realised Adam. Furthermore, if these are possibilities entailed by functionalism, then the following is likewise a possibility entailed by functionalism:

How could God¹ and God² ‘engage’ each other in ways that are capable of realising another instance of the mind of God¹? One way would be to do it in the way it’s done in the case of the Socially Realised Adam. A more traditional trinitarian approach could lean into the love motif noted in the introduction to this section. Perhaps God¹ and God² are drawn to each in love and decide to become a we, continuing their existence and activities in a profoundly close degree of communion and enjoyment of each other. Perhaps this would move them to participate moment-to-moment in infinitely many mental communicative acts that involve joint deliberation, joint preference ordering/negotiation, joint celebration of things worth celebrating, and mutual enjoyment of each other’s company. The result of this activity could be such a profoundly complex performative agent that it actually brings about the realisation of another individual instance of God¹’s person-type.Footnote ²² There are different stories one could tell. But here’s the main point: if functionalism entails that individual persons can be produced through the activity of individual persons, then it follows that it is possible for God¹ and God² to intentionally engage each other in such a way that their joint activity re-realises God¹’s person-type once again.Footnote ²³

Now, God¹ and God² are realised in the same divine system, so if the joint activity of God¹ and God² that yields God³ were also to take place completely within that one divine system, then God³ would himself be realised in that same system. Furthermore, God³ would arguably inherit his knowledge, power, and rationality from God¹ and God². Recall that God³, on this view, is a re-realisation of the mind of God¹ that results from the mutual activity of God¹ and God². This implies that when God³ decides to act or plan or communicate, that will be a manifestation of the joint activity of God¹ and God² deciding to do so. In which case, God¹ and God² will also throw their power behind that decision. At the very least it is a logical possibility.

This explanation for the existence of God³ raises a concern: if God³ is a distinct person, what is to stop God¹ and God² from engaging God³ in ways that realise a fourth divine person? There are at least two ways to address this issue. First, it could be a contingent matter that God¹ and God² do not engage God³ in this way. In which case, the fact that God is tripersonal would be contingent. This is fully compatible with Weak Functionalist Trinitarianism.

But there is another way to address this issue. Take a simple analogy involving a very small corporate board, administrative group, or leadership team. Such groups may form performative (/collective, /group) persons. While they are far too simple to plausibly yield a further individual person the analogy will remain useful. So let the Board be a performative person that is made up of exactly two individual members: Mel and Meg. Now let us ask: can Mel (or Meg) interact with the Board? For example, can Mel interact with the group of which she is a part by bringing proposals to the Board for consideration, by arguing with the Board’s past views, by challenging the Board’s future plans, and so forth? Yes, obviously. But in order to do so Mel must either interact with herself or with Meg. But when Mel interacts with her group in this way, she does not bring forth any further performative persons even though she is interacting with the Board.

Now, the magnitude and variety of interactions between God¹ and God² that would be needed to bring about God³ would be immense and for that reason quite different from the previous example just given. But there remains a way the divine case parallels the analogy to small performative persons. For suppose God¹ wants to, say, communicate something to God³. To do that, God¹ would have to interact with that which realises God³. But that would involve either interacting with himself or with God². But why should God¹’s interactions with himself or God² be expected to produce any additional individual beyond the individual those very kinds of interactions already produce, that is, God³? The very idea that they could is as strange as thinking that Mel’s engagement with the Board generates a further performative person. Stranger still is the idea that such interactions must produce a fourth person. For it seems at least possible that such actions fail to do so. We can state the point here as an independently attractive conditional: if an individual person x results from the complex interpersonal interactions of some other people y¹-yⁿ, and y¹-yⁿ only engage x by further interpersonal interactions with y¹-yⁿ, then it is possible that that activity does not yield any person in addition to x. Perhaps we could strengthen the consequent from a ‘possibly not’ claim to a ‘cannot’ claim. But the former is all that is needed to defend the consistency and conceivability of the functionalist account of the trinity.Footnote ²⁴

At this point we can appreciate the argument for Weak Functionalist Trinitarianism:

We have, at this point, provided evidence for a conditional resolution to the inconsistency and intelligibility aspects of the trinitarian trilemma. That is, conditional on a suitable version of functionalism being both consistent and conceivable, there is a consistent and conceivable functionalist version of trinitarian monotheism.