Last visitors [hide]
1)   admin at Sep 22, 2014 [12:43]
2)   huangyn2j at Jan 17, 2013 [15:50]
3)   KevinORegan at Dec 28, 2010 [16:43]
4)   kevintesting at Jan 01, 1970 [00:00]
5)   test3jan2011 at Jan 01, 1970 [00:00]
6)   del at Jan 01, 1970 [00:00]
7)   randybosh at Jan 01, 1970 [00:00]
8)   StephanieSpino at Jan 01, 1970 [00:00]
9)   Humberto321 at Jan 01, 1970 [00:00]
10)   Marcossp at Jan 01, 1970 [00:00]

Key Concepts 1: Cognitive and Conscious Access

The sensorimotor approach claims that experiencing a raw sensory feel consists in having cognitive access to the fact of currently exercising mastery of a sensorimotor skill. The following sections will read somewhat like a glossary, where every term in this definition, plus any necessary additional terms, are set out in detail.

In this first section we try to define what is meant by cognitive access. It should however be noted that though the notion of cognitive access is necessary in the theory, its precise definition is not critical and is to some extent a matter of choice. In order that the notion of feel be applicable not just to humans, but also to other creatures, we make the distinction between cognitive access and conscious access. This distinction remains problematical however, and there remain other difficulties, in particular in regard to the notion of 'self'. We look forward to contributions and commentaries to help clarify these problems.

Cognitive Access

We will say that having cognitive access to X is: being poised to make use of X in reasoning, rational action, planning or communication. Note that we want this notion to be applicable not only to humans, but also to animals and artificial agents. Thus we admit that the reasoning, rational action, planning and communication involved be of quite limited kind. On the other hand we use the peculiar term "poised" in order to stress that we do not simply mean that the agent is now using X in its cognitive processing. We want the agent to be in a state where it can but need not use it. For this to be possible, the agent must have sufficient complexity for it to make sense to say that it has some choice in the matter. On the other hand, note that we do not insist that the agent know that it itself exists as an entity or that it know that it has the choice or the cognitive access or the purpose in question. Thus the notion of cognitive access we are proposing is probably more primitive than what Ned Block calls "conscious access" (see next section).

As an example consider the tic tac toe playing machine. The machine is built to attain a purpose, to win the game, and at each juncture there is a choice about what move to make. The machine registers your move, evaluates the possibilities and prepares an appropriate response. Seen from the outside, it makes sense to say that the machine chooses which move to make and that it has rational cognitive activities since it can reason, plan moves in advance, judge, and communicate to a limited degree: all this even though the machine itself does not know that it has these capacities. The machine is poised to make use of your move in its further rational behavior. We must say the machine has cognitive access to your move.

But whereas the tic tac toe playing system has cognitive access to your move, it does not have cognitive access to the fact that right now it is in a sticky situation, or that in general it plays a mediocre game, because these are things that the system is simply not programmed to 'think' about: it does not make use of them in its reasoning, planning, judgment or communication.

Another point. Suppose we take the tic tac toe playing machine and connect its inputs and outputs to colorful buttons and lights which have no resemblance whatsoever to playing tic tac toe. We install the machine in an art gallery and allow people to press on the buttons and observe the resulting light display. The machine may go through exactly the same "mental states" as before, yet it would be inappropriate, seen from the outside, to ascribe the faculty of thought to the machine, or to say it has mental states, makes decisions and choices. It therefore makes no sense to say that the machine has cognitive access.

These examples show that the point of view that one takes determines whether we can apply the term "cognitive access" to a system. Cognitive access is not something that a system possesses inside itself, it is a convenient way of describing the potential behavior of the system as seen from the outside, in a particular context and from a particular point of view.

The importance of point of view is also shown by the following example. It turns out that there is a generation 2 model of our tic tac toe machine which plays exactly the same game as the generation 1 machine, except the electronics has been simplified: to save power and increase speed, the generation 2 machine uses a pre-calculated lookup table to determine how to counter each possible move. For the generation 2 machine it seems no longer to make sense to say the machine has cognitive access to your move because there is no more evaluation, deliberation or choosing of possibilities! The trouble is the two machines look the same on the outside. Does my machine have cognitive access or not?

I think that there is no fact of the matter. The notion of cognitive access is intrinsically a relative notion: relative to the 'point of view' one takes about the cognitive powers of the system under consideration. The outside observer observing the system must consider that it make sense to suppose that the machine is "deliberating", evaluating possibilities and coming to a decisiion. But ultimately we know this is just a stance we take, a matter of preference in talking about the behavior of the system under consideration, irrespective of the actual goings-on inside the machine. In this sense, cognition is in the mind of the beholder, not in the mind of the entity that is doing the cognition.
More on Cognitive Access?

Conscious Access

Ned Block uses exactly the same definition that we have proposed for cognitive access, but calls it conscious access: having conscious access to X means: being poised to make use of X in reasoning, planning or speech.

But from Block's and others' usage, it would seem that more is implicitly assumed in conscious access than in cognitive access. Conscious access of X is something like what is meant under the common usage "being aware of" or "noticing" X. Conscious access to X thus seems to include not only what we here have called cognitive access to X, but also knowledge that the agent has this access. In other words conscious access requires the notion of a "self", and requires that this self "know" that it currently has cognitive access to X. Perhaps there is a relation to Rosenthal's "higher order thoughts": conscious access seems to be: the self having cognitive access to the fact that it has cognitive access to X. Example: the tic tac toe machine does not have conscious access to your move because the machine has no sense of self, and no faculties to reason about the fact that it may or may not have cognitive access to your move.

(It is interesting to recall that we are maintaining that the notion of cognitive access is in the mind of the beholder, and not in the mind of the system doing the cognition. This would apply also to the meta-level cognitive access that we are using in the definition of conscious access. The implication of this is that conscious access is also in the mind of the beholder. In other words, the question of whether or not a system has conscious access to something is not a fact about the operation of the system itself, but it is a point of view that other systems can take about the behavior of the system. I suspect this idea is similar to Dennett's idea of the "intentional stance".)
More on Conscious Access?

The self

The self is a notion that an agent can use in its reasoning, planning, judgment and communication, which refers to the way it, as an entity, interacts with its environment, and in particular, with other entities like itself. In the case of agents living in a society of other similar agents, the notion of self is particularly important because it may be used to facilitate communication between the agents. We take the notion of self to be a non-magical notion which can be implemented in a sufficiently sophisticated information processing system.
More on The self?

Tricky examples with and without cognitive access

We shall now consider some examples of cognitive access in which we take the "normal" viewpoint about humans or animals as agents acting purposively in an environment consisting of other similar agents in a real physical world.

A human seeing a red traffic light and talking about it. There is clearly cognitive access (also, presumably, conscious access). But the situation is not so clear in the case where a human sees the red light, presses on the brake, but does not "notice" the light because he or she is attending to a conversation for example. In this case there is certainly no conscious access. But is there cognitive access? Is the person poised to make use of the red light in their rational behavior? First, should we consider the stopping at the red light as rational behavior, or was it just behavior per se? Furthermore was the person "poised" to make use of the red light in this rational behavior, or did the influence merely come about automatically with no possibility of another outcome?

The TV viewer who is exposed to a subliminal stimulus: The stimulus influences his rational cognitive processing because it causes him to vote for George Bush. But should one say that the viewer is poised to make use of the stimulus? It's not clear whether or not there is some unconscious mental operation going on which results in a decision about whom to vote for, or whether ultimately the viewer has no choice about the matter...

Blindsight patients who are not able to "see" the stimulus in their blind field, even though they are able to manually grasp the object if exhorted to do so, do not have conscious access to it. Do they have cognitive access?

A dog seeing a cat and running after it? This is not completely clear. Does the dog just automatically run after the cat, or is there some kind of rational planning involved? Does the dog have a choice? If both these conditions are fulfilled, then it makes sense to say it has cognitive access to the cat. Is there conscious access? Presumably dogs have a less developed notion of "self" than humans. If they have such lesser "selves" then they have correspondingly lesser conscious access.

A fly seeing another fly and chasing it. Here we would probably begin to say that since there is little choice and little cognitive and rational planning there is no cognitive access.

A country registering another country's war declaration: At the scale of countries it seems to make sense to say that there is reasoning, planning and communication. Thus here it would seem that the country has cognitive access of the war declaration.

Furthermore countries clearly have a sense of self, or at least identity, and so under our definition, it might even make sense to say that countries have conscious access. But both for cognitive and for conscious access there might be a slight difference as compared with humans, owing to the fact that whereas humans have essentially only a single channel of communication with other humans, countries express themselves via a plethora of different media. Also, perhaps countries do not have single "selves" like humans, but have composite or articulated selves, somewhat like multiple personalities in humans...
More on Examples with and without cognitive access?

Contributors to this page: KevinORegan and admin .
Page last modified on Friday 21 of April, 2006 [08:37:02 UTC] by KevinORegan.

Free Hit Counter