Against the Computer Model of the Mind: Can We Reduce All Properties to Quantities?

Source: Ghost in The Shell.

I. Strong and Weak Representationalism

Despite Western thought’s persistence to the effect, there are nevertheless reasons to doubt the idea that the mind comes in contact with the world through the mediation of a representation. For that matter, there are also good reasons to doubt the idea that the mind is in some way separate from the world– or that it is made of a different kind of substance.

The first idea, that the mind relates to the world via the mediation of representation is in some sense necessitated by the second idea. If the world is made of something different from the mind, then there is some uncertainty regarding the relations that the world and the mind bear towards each other, especially in terms of causation.

Representationalism, which is the idea that the world relates to the world in some cases via the mediation of representations, allows mental entities to in some way relate to entities ‘out there’ in the world. On this picture, non-representational mental contents relate to the world via relation to representations in the mind. The representations themselves are the way that mental contents gain transitive access to entities in the world. Entity x has transitive access to z if x and z either relate or are related to by some intermediary entity y.

There are two types of representationalism that are worth discussing. The first form is the strong form, where the mind relates to the world solely via the mediation of representations. The second is the weaker form, where the mind bears some relations to the world that are mediated via representation, but that it also bears some direct relations.

The problem with weak representationalism is that it is by no means clear how the mind might bear direct relation to the world. If we adopt the stance that uncertainty about the hows or the whys of some proposition motivates rejecting it, then we are unable to defensibly hold the weak representational position. When a proposition paints a picture that is uncertain or unclear in one or more features, and we say this uncertainty motivates us to reject the proposition, then we have rejected it on the grounds of the uncertainty criterion.

Is this condition similar to Descartes method of doubt? It is true that the picture painted by the weak representationalist claim is neither ‘clear’ nor ‘distinct.’ However, if we opt for strong representationalism, then we rapidly find ourselves in Cartesian sceptical territory. If we rejected weak representationalism on the grounds that its own representation of the direct relation between mental and physical entities was uncertain, where does that leave us on the nature of the representations that mediate on either side of the representationalist fence?

If we reject weak representationalism, then representations cannot be physical entities– this would imply that mental entities could relate without mediation to physical entities. Therefore, we need to make representations into a special class of mental entity for the strong representationalist position to remain tenable. They must be special, because they must bear some feature that enables them to directly relate to physical entities without violating the condition imposed by the strong representational claim.

It is not immediately clear what this special condition could be. One possibility is that representations of physical entities relate to the physical entities themselves through some form of resemblance. This would allow the representations to relate directly to the physical world in a way that is directly explicable, and which therefore does not violate the uncertainty criterion.

While it is by no means clear how the representations come to acquire their resemblance to physical entities, this is not strictly necessary in order to accept the claim that they do resemble physical entities. This move is motivated by the vacuum of any other suitable explanation for the relation between representations and physical entities.

II. The Mode of Representation.

Once we have accepted strong representationalism, and once we have accepted that strong representationalism can be explained in terms of mental entities called representations that bear a resemblance relation to the physical entities that they are supposed to represent, there is still an explanation that we must provide, which is the explanation for how that resemblance is bourne.

For strong representationalism, there is only one option, which is that mental entities resemble physical entities is that the mental representations describe similar ratios between extended magnitudes given by the physical entity that they are supposed to represent. This is why Descartes– the premier strong representationalist– referred physical entities as res extensa: the extended substance.

This means that mental entity x is resembles a physical entity y if the ratios of various quantities described by entity x are congruent with the ratios of various quantities given by physical entity y. This condition is quite strict, so we’ll say that x is a representation of y to a greater or lesser degree if x resembles y to a greater or lesser degree. This also entails that mental representations are the sorts of entities that describe sets of quantities.

It also entails that of all the properties a mental representation could describe, they are either explicable wholly in terms of quantity or else are not representative of the physical entities themselves. This leaves us in a potentially precarious position with properties such as color or taste, which are not obviously explicable in terms of extension in the same way as things like speed, texture, or pliability. It leaves us in a much more precarious position with respect to another certain set of properties.

Before describing the nature of that precarity, and the nature of the properties that we are describing so precariously, it will be necessary to describe another important feature of the Cartesian picture.

III. Entities as Independent.

If the res extensa is the entity whose properties are such that they are determined solely in terms of sets of quantities, then it is also true that every single one of its properties is intelligible without reference to any other property or quantity– each of its properties must be intelligible in isolation. To say that a property is intelligible in isolation is the same thing as saying that a mental entity could represent only that property, and no other property in one instance.

This means that for any given instance of res extensa, any singular property it bears can be represented by a mental entity without any other property borne by the physical entity needing to be represented.

This entails that every instance of res extensa is representable without reference to any other instance of res extensa. To say of something that it is intelligible in certain terms is the same as to say it is representable. This entails that every mental representation of a physical entity is intelligible as isolated from any other mental entity, and therefore that every mental representation is independent from any other mental entity.

To say that x is independent from y is to say that x would not change in itself if y were to change. X changes in itself iff one of its identity-determining properties changes. A property P is identity determining if some entity is an instance of type x iff it bears property P (1).

On the Cartesian picture, all identity determining properties are non-relational. If they were relational, then they would not be intelligible independently, and therefore could not be properties of a representation.

IV. Heidegger’s Critique of the Uncertainty Criterion.

Heidegger’s critique of Descartes’ as found in Sein und Zeit can offer us at least three good responses to the account of mental representation given in sections (I – III). One response is that it may be unwarranted to assume the uncertainty criterion. Another response is that it does not allow for the mind to represent entities that bear identity determining properties as given by the relationships that those entities bear to other entities. The third response, which builds on the first two, is that it is not clear how evaluative properties might be reduced purely to quantitative ones.

The second and third responses will need us to lay more groundwork. As such, we’ll just cover the first response in this section.

Regarding Heidegger’s response to the uncertainty criterion, we might make the following argument: in normal day to day life, we assume a mode of relating to our thoughts and experiences that seeks to deal with them on their own terms, and which does not seek to take a special perspective on them. Heidegger suggests that instead of taking our experiences as something to be interrogated beyond their context, they should instead be understood in terms of their ‘average everyday-ness.’

The fact of the matter is that you and I both manage to get along just fine in our day-to-day lives without the certainty that Descartes demands. While it might seem intuitively true that a more clinical perspective– such as the one implied by adopting the uncertainty criterion– might in some sense allow us a privileged understanding of the phenomena we seek to understand through it, it is also true that in day-to-day life, we are forced to accept as true propositions about which we are not suitably certain, and whose propositional content we cannot fully explain.

If we were to take the uncertainty criterion with complete seriousness, everyday life would not be possible– living simply requires us to believe too many uncertain claims in order to function on a daily basis.The assumption of strong representationalism over weak representationalism solely on the grounds of the uncertainty principle may not be as defensible as it seem.

V. Being-in-the-World.

Heidegger’s analysis also affords us a response to the Cartesian picture inasmuch as it requires the identity determining properties to be relationally independent. If there are some entities whose identities are such that they are not relationally independent, then we have reasons to reject the Cartesian picture.

One example of an entity which may not be relationally independent is a hill or a mountain; or a slope, more generally. To an ant, a hill which is easily surmountable within an hour’s walk for a human being is not easily surmountable. As such, given a particular slope, humans might qualify as the sort of beings which could easily surmount it while ants might not. If ease of surmountability is an identity condition for a given slope, which it may be, or if ease of surmounting a given slope is an identity condition for the type of being which surmounts slopes, then we have a reason to reject the Cartesian picture on the grounds that its independence condition cannot be satisfied.

Furthermore, if ease of surmountability is an identity determining property for entities such as slopes, then we may have an example of a case whereby the qualitative reduction condition cannot be met. In the case of a slope of certain length and incline, while we certainly can analyse the properties of the slope in terms of numerical extension and angle from horizontal, these properties tell us nothing about its surmountability unless we reference the sort of being that can or cannot surmount it.

While it may be possible in theory to offer a relational analysis of the extended properties of the mountain in relation to the extended properties of whichever being is proposed as possibly surmounting it, this would likely require deepening the base of knowledge collected via the behavioural sciences to a degree completely precluded by practicality. While it might be possible in theory to quantify the conditions required for evaluative properties like ease to obtain across cases, it is by no means clear how this could be done. As such, claims in favour of a quantative reduction can be suspended as premature, at least.

Footnotes.

(1). In this sentence, I use the abbreviation ‘iff’ twice– this is not a typo. In analytic philosophy, ‘iff‘ is often written in place of ‘if and only if‘: ‘x iff y‘ means if y then x and if x then y.

Bibliography.

Descartes, R., & Cottingham, J. (1986). Meditations on first philosophy: With selections from the Objections and Replies. Cambridge [Cambridgeshire: Cambridge University Press.

Heidegger, M., Macquarrie, J., & Robinson, E. (1962). Being and time. Malden, MA: Blackwell.

Further Reading.

Clark, A., 1997, Being There: Putting Brain, Body and World Together Again, Cambridge, MA: The MIT Press. (Cbapter 8)

3 thoughts on “Against the Computer Model of the Mind: Can We Reduce All Properties to Quantities?”

  1. I like this post. Presently I’m reading Descartes’ meditations on the first philosophy and I’m interested in what he says, although I do not endorse some of his claims. I’d like to ask you, does representationalism necessarily entail either substance or property dualism according to your perspectives?

    Like

    1. I think representationalism probably implies dualism, because representationalism is the claim that the mind is or contains representations of something else. I suppose you could get a really weird, boring representationalism if you were an idealist where only mental contents and representations of those contents are real, but I don’t know what explanatory power that account would have, and it seems a little overcomplicated as compared to a straight idealism without representations. You could only have this account with a weak representationalism, where the mind is only partially representational. If you go for idealism and strong representationalism, then you get representations all the way down and I think it’s fair to say that you can only have representations if you have something represented.

      Additionally, in the idealism with weak representationalism, so where there is both representational mental content and non-representational mental content, we’d need to ask whether the mind could be conscious/aware of the non-representational mental contents. If it could be aware of them, then we’re left wondering why this account would need both representational and non-representational mental content. There’s no explanatory power there. But if it couldn’t be aware of the non-representational mental content, then we’re in a situation that to all intents and purposes looks like dualism, where there are mental states that represent other states which are not immediately accessible to consciousness– it is unclear what the difference between material substance and this unrepresented mental substance would be.

      So that’s idealistic representationalism off the table. On the other hand we have materialistic representationalism. This would require that we allow material states of affairs to represent other material states of affairs, which I don’t think is impossible, though I don’t know exactly how it would work. I don’t feel comfortable arguing against this horn of the dilemma, given that I don’t feel as up to date as I’d need to be in the philosophy of mind around strict materialism. I mean, theoretically something like DNA and the cell division cycle could entail materialistic representationalism on some level, but I’d also be tempted to ask whether that just seems like representation to a person observing it.

      One of the better definitions of representation– and I can’t remember if I put this in the essay as I’m commenting without looking at it right now– is that x represents y if x can replace y in a sentence or a space without a loss of meaning. I don’t know if you’re a fan of that definition, but it does seem like we need to introduce a concept of abstract entities like meanings for that to make sense.

      While I suppose it would be possible to make representationalism about the mind compatible with materialism, I do think you’d have to end up in a really odd position, and intuitively I don’t think there’s much value to be found pursuing that track, though I am open to being shown that I’m wrong.

      I suppose you could opt for neutral substance monism, but your best option there looks like your best option for the idealism case above, I think, so again I don’t see where it would look different from substance dualism.

      I think simply the face that representationalism involves representations, and that there seems to be this intuitive need for representations to span more than one modality to be a real representation puts us in a situation where we need at least two types of stuff we can represent across.

      So yeah, I’d suppose that on this view of representationalism, you’d need dualism of some sort.

      Liked by 1 person

  2. I see, I don’t know a lot just yet but I’m actually reading Heil’s introduction to philosophy of mind and Bermúdez’s introduction to cognitive science. I don’t feel like Fodor’s LOT and Simon and Newell’s physical symbol system hypothesis would actually require some form of either substance or property dualism. I’m not a representationalist, but I don’t see how we need dualism for representationalism to be successful, honestly. Probably I haven’t read enough, but I’ll try to write something about it when I finish my readings on Simon and Newell, and Fodor.
    Thanks for your answer! 🙂

    Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s