The idea that memory is some sort of storage room is an ancient one. Plato and Aristotle had developed a forerunner of this theory. They claimed that memory is a wax tablet on which the things that we remember make an impression.
There are a couple of advantages to the idea that our memory is some sort of storage room or archive. The foremost advantage is that this idea provides us with an answer to the following question: where are the things that one remembered stored? The thing (words, faces) that one remembers have to be stored somewhere, otherwise we cannot use them. If someone does not except the storage room/archive notion of memory, he/she will have provide us with a different explanation.
What is stored in one’s memory cannot itself be remembered, for otherwise I would have actual physical chairs stored in my memory. The object that is stored in my memory is a representation of the thing that I remember. This means that there is an ambiguity in the phrase ‘object of a memory’. Both the chair and the representation of the chair are objects of my memory. But my memory must be about the chair. It is the chair, and not my representation of that chair, that is brown.
There are three possible answers that might enable us to avoid this difficulty. Either the memory is about the chair and the representation of the chair plays no role, or the memory is of the representation only, or the representation mediates between the subject of the memory and the object of the memory. The first option, that the memory of the chair is of the physical chair, and not of the representation of the chair deal with this objection, but it leaves he original problem, how we are to account for the fact that our memories must come from somewhere if we are to use them, is left unexplained. The second option is even worse. According to this theory my memories about a girl are not about her, but about my representation of her. This means that, according to this theory, memory is not really about the world. This means that proponent of this theory cannot, at least not immediately, explain why we are capable of using our memory in action. This option, that memory is about the world, and refers via a representation, tries to deal with this difficulty. This is, however, not easy. The proponents of this theory have to explain how, and in what way, our representations are, linked to the subject and to the object. In the following section, in which I will start by talking about conceptual and propositional memory only, I will give an account of the notion of representation.
1.2. Representation
What is a representation? And what does it mean to say that a representation is about something? There are several answers. The first one is that a representation is a picture. So my memory of a chair is a picture of that chair that is somehow connected to the physical chair. According to this theory it is connected to the chair, because it pictures it. Now we have to explain how this is done. A picture is normally connected to that of which it is a picture of, because the elements in the picture have a likeness to what is represented. This stands in need of an observer, someone who sees the picture and the chair. For in order to use the picture as a way of remembering the chair, I have to be able to compare it with the physical chair. This means that I have to remember how to use a picture. And how do I remember this? Via another representation? This would mean that I remember how to use the picture via another picture. Of course I also have to remember how to use this picture, so I need another picture. In other words, this theory dissolves into an infinite regress. But it gets worse. In this comparison I compared the representation to a physical picture. But our representations are mental. This means that we need a mental observer, a homunculus. This homunculus must itself have a memory, (of how to compare them) to enable it to compare the picture and the chair. This is done via another picture and by either the same or a different homunculus. If it is the same homunculus, we still end up in an infinite regression of pictures, if it is another, second, homunculus we and up with an infinite regression of homunculi.
Can we evade this difficulty by saying that it is not a homunculus, but me that compares the picture with reality? No. I will till get into the infinite regress of pictures. It also means that I attribute cognitive capacities to one part of my cognition, and via this route we get back to the homunculus problem.
What we need is an account of what representation is without a homunculus.
Another way of looking at it is by saying that a memory of something consists of an information loop or of a recursive function.[1] The difficulty with this view is that one either sees information in the normal sense, as a sequence of symbol with a meaning, or just as neurons that are firing. The first possibility brings us back to the homunculus. For the information has to be read, heard, seen, and understood. But this means that our explanation of memory presupposes an entity that already has memory. The other possibility is that memory is purely physical. In this case we have to explain the relation between the physical, causal world and the normative world. For memory has a normative dimension.
Until now I have only discussed propositional and conceptual memory. But most memories are much more complicated. A memory of a chair involves both the concept ‘chair’ and the actual chair. This means that I must be able to remember how the word ‘chair’ is related to the chair. And when I remember a chair, my memory can revolve mainly about the word, but it might also consist of images of the chair. And when I think about the chair in images and when I see the actual chair, I am immediately able to connect the word ‘chair’ with it. This means that there is a connection between the conceptual memory, the image memory and the physical chair. The meaning of the concept ‘chair’ must be, if they are to connected in the use of my memory of the chair, related to the chair itself and the memory of the chair must involve the meaning of the concept ‘chair’. The structure of factual memory will be discussed later, in section 1.4.
Meaning is normative. Perhaps something can be said for the idea that remembering a concept consists in our being able to use the concept in the appropriate conditions. According to this analysis memory consists of a set of abilities, namely the ability to use and explain a concept. Of course the question if an ability to use a concept presupposes a memory of the concept. Must the concept that I am to use be stored somewhere? Perhaps not. What if one were to say that the idea that a memory must be stored is connected with the idea that a memory consists of a representation that, in order to be stored, must have some sort of physical dimension? The information that is stored in the library consists of books. This gets us back to the homunculus. But what if the information is stored in the sense in which information is stored in a computer? Well the trouble with this idea is that one, when one uses this theory, has to say that information is physical. Of course the books are physical as well, but they only contain information, because we can read the sentence in them. But how is one to read the information in the brain? This can only be the case if the brain structures that are involved can be used as norms.[2] What if a memory cannot be as spatio-temporal item because of the homunculus fallacy? This would mean that the idea that memory consists of a storage room or archive is false.[3]
Nevertheless, we still need to understand how it can be that our memories are, in a given situation, immediately available. In order to do this I need to say something about what the structure of memory is. Memories can be of fact, and about how to do something. The first sort of memory, factual memory, consists of conceptual memory, propositional memory, memory of what things look like, and mental images. In section 1.4 I will explain the structure of factual memory. Often a memory has both a conceptual and a non-conceptual component. Before I can get into this, I will first have to explain what concepts are and how they are connected to reality.
1.3. Concepts and reality
How is a concept related to reality? What enables us to use a concept to refer to something in reality? Of course a concept is internally related to the thing that it stands for, but this does not solve the difficulty. For in order to use a concept to refer to something, we must first identify the thing that we are to refer to. Presumably via its properties. But then, either the concept has sub-concepts to refer to the properties, or the concept is used to refer to the thing with its properties. In either case, we must still be able to identify the properties of the thing, and connect the concept with it. I can use the concept ‘chair’ to refer to chair a and chair b, because a and b share a number of properties that enable me to identify them as chairs. The concept refers to any and all things that have this set of properties. So the concept chair stands for a set of things that have a series of properties in common. But if this is right, I would have to be able to identify or recognize the properties in order for me to use the concept. These properties have to be identified, at some fundamental level, non-linguistically. Otherwise one would get into an infinite regress. I would identify an object via its properties, and its properties via other properties… Of course I can name the various properties of the objects that I can refer to, but I am able to refer to the properties a well. This means that there must be some level at which I can grasp a property, that does not require my grasping any further properties. Besides, in the learning process one learns the concepts for some things before one learns property-concepts. Nevertheless, the learner must have some grasp of the properties of the thing to which he/she refer to, otherwise we cannot understand how he/she is able to use one concept (say chair) to refer to many different objects (the members of the set of chairs). Does this mean that reference is based on a non-conceptual grasp of properties via which one identifies the object that one is to refer to? The grasp must be non-conceptual, for otherwise an infant could not learn referring terms.
Of course this does not solve the problem. What is it that we are actually looking for? Ought to reference to be explained in causal terms? Could we for instance ay that reference is just that the data from the senses are neurally connected with the auditive and vocalized data? The concept, in this case a neural network, remains preserved. The next time I either use or hear the word, the network of the word I connected with the network of the visual experience, the non-conceptual properties, and therefore I am able to use the concept to refer to the entire individual of the set. If I see a chair, my chair neural-network I trigger, which is connected with my chair network-concept. Or if I see a chair I can immediately call it chair. So the learning process, the recognizing of what the word stands for when someone uses it and the actual use of the word in appropriate circumstances can be accounted for. But there is a difficulty. For if reference is causally explainable, then a certain set of norms is causally explainable, and if norms are explained causally, then their normativity is gone. For then it is also possible that someone uses the word ‘table’ to refer to a knife cannot be corrected. It is just another reaction of a physical system, as our response is, the argument for the causality of norms is, just like this sentence.But that cannot be the case, for sentences have truth-values.
Ludwig Kamphausen
[1] Pieter Seuren suggested this account of memory to me in conversation.
[2] This argument is basically the same as the one by P.M.S. Hacker in Wittgenstein: meaning and Mind (Oxford
[3] Strictly speaking it is nonsensical. If something is false, it could be true, but given the objections, the idea that memory is a storage room cannot be right.
Geen opmerkingen:
Een reactie posten