The point is forming an inner representation from the photo doesn't add any information to what is already in the photo. An inner representation is just as static and dead as the photo.

Take this representation: ------->

You take that as an arrow pointing to the right, and perhaps if you were driving, you would respond by turning to the right to follow the curve. But the representation itself doesn't explain anything. Turning or looking rightward when encountering that representation is convention. We could just as well have learned to turn and look to the left. The representation, then, only comes to life when functioning in a wider context (in this example, the wider context is human action).

Computers sometimes represent red as the state represented by the RGB triplet (255, 0, 0). Let's say that by lucky coincidence the brain also represented red with (255, 0, 0) and green with (0, 255, 0). But what do the representations tell us? We couldn't discover the difference between red and green by examining those two triplets. We would have to expand the context. And when we do, it would be clear that the representation itself was arbitrary. It was how the system behaved that mattered.

You hand me a stack of punched cards that are supposed to allow me to do my taxes. So I put them in a Jacquard Loom and get nothing but part of a ratty rug and a jammed-up loom. The cards are representations, but without the surrounding context of action, they have no inherent meaning.

A representation doesn't explain anything.

But if a pie cannot tell you what a pie is, a representation surely cannot either. A representation of a pie contains no information that the pie doesn't already present.

What do representations really explain?

Representations explain nothing; they represent things.
The concept of representations explains many things.

Grant Hutchison

In most situations, I would agree, but then there's that schizophrenic person who we know is hallucinating, who we might want to think phenomenalistically about, and the engineer trying to get the best mixture, who might think entirely realistically. In neither case do we need any bridges between what is real and what is being pictured, so we can just do without the representationalist overhead. Someone using representationalism as an epistemology, rather than a meta-epistemology, would be somewhat at a loss in either situation. My meta-epistemology is "tailor your epistemology to what answers the questions you need answered", and don't worry about blanket justifications. It works great for things like quantum mechanics interpretations, for example, and I should think perhaps also understanding perception and consciousness.

__________________
If we understood everything going on in the head of a pin... we still wouldn't know not to step on the pointy end.

People think the problem with models is that they are limited by our minds, but the greater problem is that our minds are limited by our models.

Why would I need to form a representation if I already have one, why does static=dead, and why would I want a dynamic representation of a static entity?
It matters not when a representation "comes to life", only that it does. That is the purpose of them, to be available for whatever is needed from them.
You are again confusing representations with labels. A computer can label something like that, but that is not forming a representation. If a representation of red were that easy to accomplish, we would hardly need a brain to do it.
Labels are arbitrary. Representations are not. Getting back to linguistics, the words are just labels and are arbitrary, but what they represent is the real language. Language is agreed-on connections between labels and representations, so the representations can be conjured by calling the labels. It is like the difference between a subroutine and its name-- the name is arbitrary, the subroutine is not.
Again, the cards are just labels. The representations are what they do when you put them in the machine.
It isn't an explanation, it is the very stuff of communication.Again, why should it?
Whether or not I want to bite into a pie. What else explains that?

__________________
If we understood everything going on in the head of a pin... we still wouldn't know not to step on the pointy end.

People think the problem with models is that they are limited by our minds, but the greater problem is that our minds are limited by our models.

If anyone is interested, here's a recent TED Talk by a brain scientist who had the fortune (?) to study her own stroke!

__________________
"Where the telescope ends, the microscope begins. Which of the two has the greater view?" - Hugo

"Men occasionally stumble over the truth, but most pick themselves up and hurry off as if nothing had happened." - Churchill

The purpose of my post was to examine examples of representations so that we could see their role and to understand how you think they empower us. If I take a (Polaroid) photo of the Empire State Building, the photo in my hand does not provide me with any more knowledge or skills than I already have when looking at the building.

The representation is being offered as the explanation for why you can recognize things, whether you should bite into a pie, and so forth. A representation is something that stands for something else. The issue is identifying the means by which "something that stands for something else" enables skilled or competent action.

It will help if you can list some examples of representations. Clarification of terms will reduce the chance of us talking past one another.

I take things like photos, paintings, blueprints, small plastic ships on a war operations room plotting table, and so forth as representations, that is, as things that stand for something else. You appear to be taking anything tangible or otherwise identifiable as a "label" that when present conjures up a "representation." Is there anything laying around the house or the office that you think is such a representation that could be conjured up by a label?

The representations, then, are the behaviors of the system. Is that right? If so, then the widening of context from just the "label" to the behavior is positive step towards achieving common ground here. The best way to understand a pawn is not to study just the pawn piece itself, but to see it in the context of action in an active chess game.

You seem to be using the term "representation" to mean "whatever it is that allows me to do what I do." Such a notion cannot be wrong because it really doesn't say too much.

A ship approaches an enemy ship on the war room plotting table. An officer sees this and springs into action ordering the ship to attack. An explanation: in the officer's head, a ship representation triggers the order-attack representation.

The nature of the "Representationalist" explanation is to take what is seen and stick it in the head as if inside processes mirroring outside processes explain anything. That is the essence of the diagram at the top of:

The Representationalism Web Site

It's not so much the little man in the head that should bother us. It's that what is in the head is pretty much a replica of what is outside the head with the phenomena to be explained such as color added to the inside version. It works well enough as a folk explanation. We can't go through life without explanations for everything we see, no matter how complex. But the folk explanations should slowly evaporate under the illumination provided by the disciplined study of philosophy and cognitive science.

There may be some utility in labeling my view of perception as "direct," but my views don't fit into any of the traditional categories. The Wikipedia article on direct realism starts: "Direct realism is a theory of perception that claims that the senses provide us with direct awareness of the external world. In contrast, indirect realism and representationalism claim that we are directly aware only of internal representations of the external world. Idealism, on the other hand, asserts that no world exists apart from mind-dependent ideas.

These all have the inherent model of:

world ---> senses ---> my perception or awareness

The traditional bickering among philosophers is over whether the "senses" part provide full, partial, or no access to the world.
I disagree that there are such things as "senses" that stand apart and "provide me with awareness." What might those be in regards to vision: the eyes? V1? V2? V4?

Instead, I think what is fundamental and the substrate of all we can talk about on this subject is life itself being lived in the world. That is, us grinding through our daily lives trying to do the best we can, to enjoy it the best we can.

Take a tennis player returning a serve. That's the "substrate" here. To talk of vision is simply to talk of how available optical information was taken advantage of to successfully return the serve. She has to keep her eyes on the ball and continually position her body to be at the proper location by time the ball reaches her. She has to swing her racket at a particular angle and at a particular speed. There are all sorts of feedback loops comprised of her, the racket, the ball, and the light at work here. They are "guidance systems" that guide on optical information. (By "information" here, I mean "differences" however those may exist in the environment.)

It doesn't divide up cleanly into a "senses" and "faculty of awareness" or whatever. The eyes, brain, muscles, light, and ball are all part of a single physical system. To talk of "senses" is to talk of aspects of her achievement and not to talk of specific organs.

We can as well talk about other aspects of the serve-return as the proprioceptive system or the haptic system, and so forth.

Consider the dynamics in our tennis player serve-return example. Quadrillions of "states" occurred. None of them are inherently more relevant than the others. There is no one "brain state" that will explain the matter to us.

Why does the tennis player have to train for years to do her job well? Does saying something along the lines of "she forms the right brain states" really say much? Does she learn to visually take in the arena and "tile together" everything into a single inner brain state, for example? In the wider, dynamic context, on the other hand, we may discover that repetitive practice conditions her to track the ball with the fovea of her eyes so that (1) she does not have to subtract out irrelevant background information, and (2) so that the retino-neuro-muscular system has a clean signal to track on in guiding the body to where the ball is going to be. In the dynamic view, that our vision is so gappy and incomplete is suddenly seen as an advantage, as a strategy nature employs to make the most of the resources available.

It's in the dynamic view where we seem to get our first foothold on the topic of perception.

Not "why", how.That is the issue after one has already identified the importance of the concept. Then the hard work can begin.
Representations are pretty much the projection of reality onto the ways we conceive of reality, soo come in many forms. The meaning of every word in this post is a representation, and if your representation differs from mine, even subtly, the communication is altered. A photograph is not a representation, it is also something in external reality, but a mental image is. If I can recognize someone, I must have a pre-existing representation of them. What we call familiarity is not the sum of our experiences, it is the sum of our representations.
I imagine three non-overlapping sets: tangibles, labels, and representations. Nothing that is tangible is a representation, but representations are the way we conceive of the tangibles, and labels are how we categorize them efficiently.
No, only a witch can "conjure up" something laying around the house. Our minds can conjure up representations derived from things lying around the house, and give them labels.

No, behaviors are something different, defined and accessed in other ways.
No, that is the best way to understand how a pawn behaves. (And we form other representations for that too.) It all depends on the question being asked-- someone could be enthralled with pawns, study their history and adorn their house with them, and have no idea how to play chess.

No, that would be silly. See above for the definition I'm using.
No wonder you don't like the concept of representation-- you are missing the whole point of what a representation is! A "replica" of what is outside does not "fit" in your head, like a ship in a bottle. A representation is not a replica, that would fail its entire purpose of being something that does "fit" inside our head.

__________________
If we understood everything going on in the head of a pin... we still wouldn't know not to step on the pointy end.

People think the problem with models is that they are limited by our minds, but the greater problem is that our minds are limited by our models.

Well, Wikipedia's not the ideal first stop for these things. The essence of direct realism is that we perceive the outer world directly, undistorted by subjective entities generated in our heads. "The senses" are just the most obvious way we might achieve that perception, but their function isn't particularly clear in direct realism.

All of it. From retina to cortex.

And yet you earlier seemed quite happy with the idea of training some neural network inside the head to recognize Halle Berry. So from the above, do I take it that you model the whole neural content of the tennis player's head as some elaborate neural network, responding in complex ways to the "optical information" in the world?

No-one claims such a thing.

Your "clean system" neglects the position of her opponent, the net, and the lines of the court, all of which are vital considerations. "Gappy and incomplete" become problematic, again.

We've had our first, second and subsequent footholds on perception for many years.

Grant Hutchison