Motivations in Strong AI

Say that I succeed in making a perfectly human robot, and that she then goes off and refuses to do what I tell her to do. I’d then just reprogram her motivation module so that she acted out of the motivations I wanted her to have. But this is only an appropriate response because her motivation module is not something that she possesses by herself, but a mere conduit for what I want her to do. She has no motivations, just extensions and means to execute my motivations. Such a machine might pass the Turing test with flying colors, but it would still have no self.

 

Advertisements

Analogy from creatures to God

The analogy from creatures to God happens because human beings spontaneously form descriptions of physical things which, upon closer analysis, do not seem to be appropriate descriptions of the physical things but only of a being that transcends them. Consider the Phil 101 distinction of things into “substances” and “accidents”.  At the first level of understanding, the two divide into what exists by itself and what exists only in another. Butterflies are substances, wing patterns are accidents. But the more weight we try to put on this idea of “existing by itself” the more we find that the it gets qualified, modified and exposed as limited. What we called a substance is utterly dependent on and secondary to a thousand different causes and conditions, without which it could not exist at all (think barometric pressure, temperature, air density, the activity of the sun, the bonds between molecules, the generating activity of its parents, etc.). Moreover, a metaphysical analysis of these things we call substances can recognize existence in them only in an instrumental sense: contingent things, for example, are defined by a reference to non-existence; and even the things that we find in nature that are not physically contingent (like mass-energy or space) nevertheless only have a hypothetical necessity. The best we can say about them is that if they exist, then they must continue to do so.

And so the idea we have of substance as what exists by itself, simple as it is, is not best verified by the thing we first learn the concept from. This does not, of course, prove that there is something higher than the thing we learn the concept from (analogy is only a way of naming things already understood) but it does open up an avenue for understanding and naming any reality that we might find beyond the physical. This is exactly what one is doing in the analogy that goes from creatures to God.

A similar sort of analysis to the one made of “substance” and “existing by itself” can be done with ideas like life, wise, good, true, permanent, duration, power, etc.

Life: It starts off meaning anything that can act by itself, or is somehow a locus of responsibility for its actions. We see this as verified in trees, though more so in animals, and most of all in human beings. And yet on closer analysis this self-activity is hedged in and qualified by a thousand different realities: subconscious drives, reflex processes, automated systems, pre-established drives, and the whole whirring colossus of what Aristotle called “nature”. What first seemed alive looks more and more like something purely natural.

Wise: One who knows the ultimate causes of things. We can get a decent view of things things in a vague and general way, but the more we try to make this precise – the more we actually try to know them – the more the nature of these things fades into dialectical definitions, large and controversial assumptions, faith in the work of others that we cannot do ourselves, the confusing jumble of contradictory or hostile opinions from experts who are way more intelligent than we are, the limitation we have to a small set of languages and traditions, etc., To know the ultimate causes of things is not the sort of thing that best belongs to the sort of knowledge we have.

Good: What satisfies the desire of the will. On closer analysis, however, we find that nothing satisfies the will, but the will in a certain respect, or at a certain time, or in a particular circumstance.

Permanent: To endure or abide. This is an odd thing to say of anything temporal, whose whole existence consists in being divided from past and future.

Power (or cause): The ability to cause change in others, or be the source of the existence of something. But none of our makings ever cause existence, but only existence in some sense (though generation does come close). Likewise, all physical changes are bi-directional, and the cause of the result is more the interaction of things than the action of one thing.

Notice that all of these analogies move from one sort of confused or imprecise idea to the notion that the idea can only be precisely or simply realized in a sphere transcending the sphere we begin with. There is also another way of making our vague ideas precise, sc. to strive to understand these ideas more exactly in the very realm where we first find them. In this way, we might use operational or testable definitions. This is simply another, non-competing way of making concepts precise.

Lecture on monothelitism and the nature of the human will

“We come to the last of the great Christological heresies today: the seventh century heresy of monothelitism (to pronounce, accent second syllable). Like the other heresies we have studied, it was not set forth by obscure cranks but by Patriarchs of the Church – in fact, this is true in spades of monothelitism, which was advanced by no fewer than three major patriarchs: Sergius, the Patriarch of Constantinople, Cyrus, the Patriarch of Alexandria, and Athanasius, the Patriarch of Antioch.

“Monothelitism takes the orthodox teaching of Christ’s hypostatic union for granted, and in fact cannot be brought into any direct contradiction with the great christian creeds. All monothelites agree that Christ is two complete natures, human and divine, in a single divine person. They simply say that there is only one divine will in Christ and no human will.

“Repress your desire to refute this opinion right away. The response to it is, to be sure, close enough at hand: i.e. there is no such thing as a complete human nature without a free will. This is the basic response to Monothelitism, and it does well enough to refute it, but it also fails to see just what is so attractive in Monothelitism, and just what is so important to see as wrong in it.

“Let me first give an argument for monothelitism:

If Christ has a human will in addition to his divine will, then he can actually choose and act to do both X and non-X.

But it is impossible to actually act and choose both X and not-X

“I stress this is a choice terminating in action, and since that action has to either happen or not,  the consequent is really impossible. We can’t set up a scenario where it would be possible for such a thing to happen.

“The response to this is, in all the literature that I read, that the argument fails to distinguish the power of will from the active exercise of the will. There is no reason whatsoever why two wills can’t will the same thing – friends do this all the time. True, there cannot be two active exercises of the will in Christ, but all the hypostatic union is positing is two powers of will.

“This response is fine, but the monothelite response gets to the heart of the matter. Let’s put it as an aporia where any two statements can be true, but not the remaining one:

a.) The human will is free

b.) A free will can choose to do something other than God’s will

c.) Christ’s will cannot choose to do something other than God’s will.

“The monothelite is simply taking (c.) from the argument that was raised against him, but this c. seems to be the weak part of the argument. Who can deny that the will is free, or that freedom means the ability to sin? Our very understanding of choice thus becomes implicated in our Christological ideas.

“As startling as it seems, however, we deny (b.). Here’s my claim the will sins not as free, but as defective. Here’s the argument:

God is both infallible judge and standard of goodness

The will deviates from the standard of goodness not as free, but as defective.

“We spoke about the minor premise when we discussed the Fourth Way. The major premise is a priori – since to be defective just means to deviate from a standard. A thing does not need to be rational to deviate from its standard (all sorts of non-rational things do so), and so there is nothing rational or free in simply deviating. True, we have special names for deviance in rational and free creatures (malice, sin, willfulness, vice, etc.) but they remain deviations, and thus corruptions of will as opposed to exercises of it.

“This complete division between a defective will and a free will is one of the things that Augustine was driving at in the climatic chapters of the climatic book of the Confessions. We do not do evil because we have an evil will, but because we have one will that wills imperfectly.

“But the Theory here has to flesh itself out in practice, and I find that it does. Consider your interior life when you are actually setting out to do something bad. You don’t, it seems to me, look at a set of “bad things” and “good things” and choose the one over the other, rather, there’s just one thing – the bad thing – that you talk yourself into forgetting or overlooking the evil of. You stop thinking about the thing as wrong, and just consider what you want in it. All evil is done in a mental oblivion or forgetfullness of the peculiar evil we are committing  – though an oblivion that we chose to induce in ourselves. Now in one sense this is the work of a free or rational will – we choose to render ourselves forgetful – but the act of the will is precisely to render the will defective, i.e. to deprive it of the very principle that it should act from. Evil is thus not so much the act of the free will, but the first act we commit after the will has ceased to be free.

“And so the Monothelite heresy points to a way in which we fundamentally misunderstand the human will; it points us to the crucial difference between a free will and a defective will. Both freedom and defectiveness are sorts of indetermination, but they are utterly opposite, and even contrary to another. If freedom is a perfection of the will, a will is less free to the extent that it is defective; it is more free to the extent that it is unable to deviate from a standard. At the limit of perfection, we find a being that is at once perfectly free and entirely unable to deviate from the standard in any way – indeed a being that is the standard. In the peculiar case of Christ, the sense of him being anable to sin is that his will was to strong and to perfect to do so.  We ultimately reject monothetism out of devotion to the manliness, the fortitude, and the intensity of focus and attention in the soul of Christ – a fortitude that we all strive to attain for ourselves as well.

An account of the hypostatic union

Accounts of the union of Christ are often made by way of contrast with heresies, and Nestorianism and Monophysitism can stand for the rest. One account of the difference can be put by way of the different resonses each of these give to the following counterfactual: If the divine person were removed from Christ, what would be the consequence to Christ’s human nature? 

For Nestorianism, this would leave the same human nature that was present beforehand, and even the same human subsistence.

For Monophysitism, this would leave the same human nature that was present beforehand as a part or element in Christ.

On a hypostatic union account, the removal of a divine person would make the very human nature itself cease to exist.

One account of monophysitism gets very close to the account of the hypostatic union, sc. saying that God is joined to man in Christ like the soul is joined to the body. On this account, there is a sense in which the removal of the divine person would mean that the human nature of Christ ceased to exist, namely in the same way that a body ceases to be properly human after death. Such an account would get us close, but not close enough. On the hypostatic union account of Christ, I think we have to say that if the divine person were taken away from Christ,  then his human nature – even his particularized human nature – would not remain. Whatever the removal of the divine person left us with, it could not be considered human.

Negating something in creation

When I make something, I impart some pattern conceived in advance to something existing substantially. The ink and the paper (or their equivalents) are already there and I just give them a new accidental arrangement; the ingredients are already on the counter and I just put them closer to each other and heat them up.

In divine creation, we keep this sense of the pattern or idea being in God but we need to negate the substantiality of things. We understand the procession of the divine idea to things, but we can’t also forget about the procession of substantiality. It is not enough to posit an intentional pre-existence of things in God: this is mere design or divine tinkering. We need to posit a physical pre-existence of things in God.

I’m not sure what I’m saying here. But this isn’t a crazy line of analysis in the Thomist tradition. It seems to be the standard way of, for example, viewing eternity in relation to things since John of St. Thomas.

Family Photo, 2013

DSC_9309

 

Family shot

Our baby girl…

 

felicity
DSC_9232 Our Oldest

DSC_9267Our Middle….

DSC_9340 Super Heroes…

DSC_9347 Super girl, chasing after daddy

Mere Christianity and the creeds

I’m fascinated by Plantinga’s account of mere Christianity as “the intersection of the great Christian creeds”. On this account, mere Christianity is fundamentally and overwhelmingly Incarnational and Trinitarian. This seems to me unavoidable even in the Apostles’s Creed, but at the very least there is a clear development in the Creeds towards an explicit Trinitarianism and traditional christology (i.e. Christ is two natures, remaining whole and diverse, united in one divine person.)

There seems to be something fundamentally right and even unavoidable about defining mere Christianity in this way, but it also seems to lead to a view of Christian argumentation and discourse that is very different from what is usually called “mere Christianity”.

1.) What we call “mere Christianity” seems more heavily weighted towards theistic proofs and the reality of sin. The Creeds, however, mention none of don’t stress this. On this account, what we call mere Christianity should be called pre-Christianity. But what is this? Paganism? The Old Covenant? Rational ethics with a splash of “religion”?

2.) We tend to argue all properly Christian claims from the viewpoint of a theologian doing scriptural exegesis, i.e. from a guy who is either working from his own wits or from some one of the approved contemporary exegetical traditions. So where are the creeds? They seem to be nowhere to be found – or, what’s worse, they are taken as objects to be critiqued in light of exegesis. The creed is simply not taken as being regulative, or as expressing a mere Christian tradition within which exegesis must be guided.

3.) We might want mere Christianity to be Apostolic or dating to the time of the New Testament writers, but it is impossible to place it there. The Apostolic age gave us everything, and so could not give us the mere basis. By definition, “mere Christianity” suggests refinement, minimalization, precision of discernment, etc, which all presupposed a large mass that was given without any such activity being performed upon it. No, if mere Christianity has any hermeneutical value, then either it comes from the Christian tradition or it starts with us – though this latter opinion is hard to distinguish from the claim that the Christian tradition has no value at all.

Essence and self in Natural theology and Trinitarianism

Natural theology and trinitarianism complement each other in the doctrine that there is an identity of essence and individual in God, since it follows from this that an account of the essence of God, while remaining formally and properly an account of the essence, would also be an account of God’s individual self. But such an essential account is not knowable to us, since it demands a third category of description transcending the abstract and the concrete, and we have no such category.

It follows that we can make no conceptual bridge whatsoever* between our account of divinity and the self belonging to this divinity. This includes the attempt to make an inference between there being only one divinity and there being only one divine self.

Trinitarianism thus highlights how negative and relative-to-creatures our idea of God and divinity is (are?), since it tells us that if we could see the essence, we would see the multiplicity of selves, whereas the only view of essence we get, like any view of essence we get, provides us with no information about the selves having that nature. Example: by looking up the definition of anything in a dictionary, we cannot tell how many things defined there are, since we can’t tell if there are any such things at all (just look up “dinosaur” or “plogiston” or “test particle” or even “triangle”).

 _______________

*This is subject to the usual restriction that we can know that God is but not what he is, which in this context means that we can if there is some divine essence, there is some divine self, but this claim is indifferent to Unitarianism/ Monism and trinitarianism. Trinitarians certainly say there is some divine self. Nothing changes about this even after a proof that the divine essence can only be one, even in conjunction with the fact that this essence is the same as the individual; for this still does not open to us the view of the essence itself.

 Briefly, the following inference is false: “God’s essence is the same as the self, and the essence is undivided, therefore he is one self”, rather, the inference is “God’s essence is the same as the individual, therefore understanding the relation between the two requires a category we do not understand”. We know nothing about the relation of essence and self in God beyond what God himself decides to tell us.

 

 

Hypostasis as opposed to particularity

We need to divide hypostasis from particularity. The words are not crucial here and can be taken as just particularity 1 and particularity 2:

1.) the particular is an instance of a universal, hypostasis is not just an instance of the universal. Mary is a particular instance of humanity, but when I love Mary I am not loving some instance of more general properties.

2.) Universals are proper to intellect, particulars to sense. But hypostasis is an intelligible reality known by experience. Experience is not the same thing as sensation, nor can it be considered a mere multiplication of sensation without a properly intellectual component. Again, particulars are mere multiplications of things and so are taken as homogeneous, hypostasis is a particular taken in its unrepeatability.

3.) Universal and particular are logical terms; hypostasis is a real term. Again, particular is opposed to universal, which is intentional; hypostasis is opposed to nature or essence which are metaphysical.

An Atheist ontological argument

Other versions of this have been given, but one that is closer to Thomistic ideas would be something like:

If God exists, then there is a definition that provides positive information that the thing defined exists.

But definitions cannot provide such information.

Therefore, God does not exist.

The consequence follows from the Platonic-Anselmian-Thomistic idea that God exists by nature or in virtue of what he is along with the idea that definition is the statement of what something is.

The argument has many precedents- it can even be read as a commentary or elaboration of Kant’s idea that existence is not a real predicate, sc. that whether something exists or not is not a piece of information that can be conveyed by an account of what it is. In this way we mihgt be able to draw a more direct line between the ontological argument (or, as here, an anti-ontological argument) and Kant’s claim that his refutation also shows the cosmological argument can’t work.

The Thomistic response is well known, though perhaps could be dwelt on at greater length: one way to articulate the response is that definitions arise from knowing intrinsic causes, but we do not know God by intrinsic causes; or that definitions are proper to finite reality whereas God is not such a thing. These responses themselves generate new conclusions – indeed, a whole science falls out of them. Another important division is between knowing that something is and knowing what it is, but this distinction has yet to be fully fleshed out by any Thomist I’m aware of.

This raises other, more interesting problems: if we do not know intrinsic causes in God, or at least what is analogous to intrinsic causes in God, what do we come to understand, as in the Fourth Way, that God exists in the same way that fire is hot?   If we update Aquinas’s example,* we would say that molecular motion is certainly an intrinsic cause of heat, and we are speaking about existence in God in the same way; so that  it is not clear how we are restricting ourselves to a knowledge of that God is and not what he is. It is not easy to see how we can predicate something positive about another first and per se wihtout giving an account of what it is as opposed to an account that it is.

____

* In Aquinas understanding of thing, heat is fire in exactly the same way that we now say heat is mean molecular motion. Aquinas was wrong, but we can update his example without losing his main point – in fact, unless we update his example we totally miss the point he was making.

« Older entries