This is another rough draft of an answer I rehearsed for last years exams and ended up using on the actual exams. This ran to four pages, hand written. I think it was the reference to the Dennett experiments at the end that clinched it good marks, on the actual exam I drew a diagram illustrating the actual results of the experiment better than I could of using only words.
"What does it mean to say that someone perceives something indirectly? When, if ever, is it correct to say that someone perceives something but perceives it only indirectly?"
This question places one immediately in one of the central debates of epistemology, that of whether perceptual data of the external world is received directly or indirectly. What exactly it might mean to receive sensory data is best understood, initially, by viewing the theory in which it is in direct opposition to.
Direct perception is a label which covers many theories, from "naive realism" to "adverbial theory". The former essentially advocates the unfiltered passage of sensory stimulation from physical objects to conscious awareness, while the later holds that what one senses are not objects at all, but the immediate sort of sensations of shape and color which allow one to make sense of the external world.
Diametrically opposed to theories of direct perception are those advocating "indirect perception". What unifies various accounts of indirect perception is the existence of "sense data", that is, intermediates between the object being viewed and the conscious perception of the subject. Sense-datum are essentially intermediaries between the physical and mental worlds, according to a dualist viewpoint, or between the external and internal, to adopt a more neutral phrasing.
Take the example of a red ball. Under a naively realistic view, seldom currently advocated, light from the ball's surface enters the subject's brain and is somehow understood as a red ball. An adverbialist would claim that the subject receiving the same sensory stimulus is viewing "redly" and "roundly", putting them together with linguistic labels to form the idea that a red ball is viewed. Under indirect theory, stimulation of the sensory organs passes to the brain, which then somehow forms an image of a red ball in the mind. It is this image, rather than the ball, which is perceived.
The implications are, of course, profound. In accepting an indirect viewpoint, one is cut off from direct knowledge of the world external to the mind. In the absence of an account of how sense data are casually related to physical objects, any sort of real knowledge becomes very difficult to justify. Indeed the very existence of physical objects may be called into question. Grice attempts to solve this problem with unified theories of causation and sense data; while the exact casual mechanism may not be understood the public nature of the sense data of physical objects seems to imply that a casual link exists.
Both viewpoints are beset with profound difficulties, for direct realism not the least being that it is simply scientifically implausible. Indirect realism, on the other hand, has other problems, such as how objects arranged in space can be arranged in a non-physical image; or how non-physical images in the mind can have color, extension and the other features exhibited by physical objects. Not the least worry is how such images might be perceived while avoiding an infinite regress.
Such objections notwithstanding, there can be little doubt that there do exist phenomenon which are viewed exclusively as images. In the case of dreams and hallucinations it cannot be otherwise, obviously enough, and in these cases in simple purposeful imagining images formed with color  and extension. Likewise, there exist cases in which it must be conceded that external objects are being indirectly experienced.
Some eminent philosophers, for instance Russell, held or hold that all objects are indirectly experienced, noting that the variations which exist in the secondary data (for instance color, texture, brightness, ect) perceived. For instance, a table might appear to be different shapes depending on the angle of the viewer and the distance relative to the table, while the patterns of light and grain of it would likewise appear different. These observations lead to the conclusion, that since the actual table is unvarying in shape and color, what is being perceived are not tables, but table images formulated in a certain way.
Such assertions may be countered an adverbalist. What is being received are merely sensations of degrees of color and shape which will vary with the relation of the viewer to the perceived object. Variations in the object are not needed to explain variations in its appearance, and neither are sense-data.
However, further support for indirect perception comes from an unlikely source, experiments related in Dennet ("Consciousness Explained"). Two dots, close together and flashing one quickly after another, are perceived as a single dot moving from side to side. If the two are different colors, say the first red and the second green, the single moving dot will appear to change color half way; in other words it seems to change to green before the mind consciously perceives the green light as having flashed at all.
Such revision of events already having occurred, in such a mundane case lacking any sort of appeal to hallucination or dreams, seems to argue strongly that indirect perception is, in fact, the normal operation of the mechanisms involved in the subject's visual perception.
In the case of the other sense, it would be reasonable to assume by analogy with vision that operation via images, that is indirect perception, is either always or at least sometimes the case. The ability of the mind to incorporate different data from the various sense is strongly indicative of indirect perception, with the received data being categorized into unified "images", allowing us to match, for instance, tactile and visual perception of the same object.
 

This isn't a question that I have seen every year on the reports, but it was asked last year and I reproduced this essay almost line for line and recieved a 73.

As promised, I finished my summer statistics class and am starting in on posting the practice essays I prepared last year. I got a 70 on this one, or one similar to it, I forget the exact question...but I think this is a very good theme for an epistemology student to rehearse, as some variant of the dreaming question appears every year for which I have seen examiner's reports. Obviously this will cover some ground similar to the one I put on here earlier, which was the very first practice answer I ever wrote, hopefully this one shows a little more maturity...
This was written in a little under an hour, and ran to six pages by hand; of course the hand-written one included many more spelling errors...
"Since there is no way of telling whether you are dreaming, there is no way of knowing anything about the external world". Discuss.
The question of whether it is possible to know whether one is dreaming or not has exercised a powerful sway in the history of epistemology, beginning with Plato and rising to become Descartes's central concern. Like other radical skeptical hypotheses, it appears to eliminate any justification on which knowledge of the external world can be based, and unlike other classic skeptical scenarios it does not entail the supposition of entities or technologies not know to exist.
According to the skeptic advocating the impossibility of knowledge, the premise (D) "It is impossible to tell whether one is dreaming", in conjunction with the traditional tripartite account of knowledge, would render any knowledge impossible. The tripartite account sets three conditions for a subject (S), to know proposition (P):
1) P must be true
2) S must believe that P
3) the belief of S that P must be justified
According the skeptic, the possibility of knowing for certain that (D) does not obtain eliminates the justification for any further beliefs. For instance, a proposition such as "I am now sitting in an exam hall" is not justified, since cannot be certain that I am not, in fact, asleep and dreaming that I am seated in an exam hall"....
A possible reply is that some non-propositional knowledge may remain untouched. For instance, though I may be dreaming at any moment, I still "know" how to walk, speak English or perform quadratic equations. However, epistemology is typically concerned with propositional knowledge, as the main source of knowledge about the world, so an acceptable response to the skeptic should concentrate on "knowledge that P" where P is a proposition. Therefore, two options are open to a thinker wishing to defend the possibility of knowledge:
A) to deny the validity of the premise
B) to deny that the premise entails the conclusion
Attempts at B have been common in 20th century philosophy. Goldman rejected what he regarded as irrelevant alternatives; Moore similarly placed the onus of proof on the skeptic, requiring that a challenge to common sense knowledge be proven, rather than that common sense itself be defended. Skeptics may reject this sort of reasoning as circular in nature-"common sense knowledge" is essentially being defended on the basis of common sense. A similar objection may be raised to Lehrer's assumption that human perception is trustworthy. Such assumptions essentially side-step, rather than address, skeptical challenges, and seem somewhat unsatisfactory to many.
An attempt to meet this hypothesis head on is Nozick's challenge to the "Principle of Closure Under Known Entailment". According to the principle, if S knows that P, and P entails Q, then S also knows that Q, since P must be true and a true antecedent cannot entail a false consequent. Therefore, if I know that P, "I am seated in an exam hall", this entails Q, "I am not at home dreaming that I am seated in an exam hall". According to the skeptic, I don't know Q, so I don't know that P, either.
Nozick replies to the skeptic by analyzing the alternate possibilities in terms of possible worlds. For instance, in the possible worlds nearest to the actual one, if I am not in the exam hall, I may be just outside the hall loitering, dealing with a broken down car, having a panic attack, etc. A vivid dream that follows consistently for such a long period as I have been writing this essay, while indeed a logical possibility, is unlikely enough not to be the case in reasonably "nearby" worlds, and therefore, according to Nozick, the Principle of Known Closure Fails when considering radical skeptical hypotheses.
However, Nozicks original argument was formulated against the "Brain in the Vat" scenerio, a hypothesis that most can agree has a very remote chance of obtaining in actuality. To moves remain to the skeptic, to deny that a very vivid dream is in fact a remote possibility; and more importantly to maintain that since knowledge is impossible given the possibility of premise D, it is impossible to meaningfully formulate the chance of such a scenario obtaining or not!
What remains is for an advocate of knowledge to attack the premise itself, which fortunately is not as difficult of a target as its long history in the philosophical literature would have it appear. Firstly, granting any accuracy to our waking perceptual knowledge, dreams with enough verisimilitude to be mistaken for  real life are extremely uncommon. This immediately reduces the skeptics challenge from a pressing concern to a sort of logical possibility that must be dealt with only out of philosophical interest. Scientific knowledge and coherentist scientific epistemologies such as that proposed by WV Qunie can operate completely unphased by the possibility of vivid dreams.
One essential point that can be raised against the skeptic is that "dreams", under any normal definition of the word, are episodic, while life is a continuous narrative. Denying this on the skeptic's part would leave us talking about very different things, without any basis for discussion. The episodic nature of dreams allows a moment's reflection to answer the question of whether one is dreaming or not, thus eliminating the skeptic's premise D.
For instance my experience of sitting here in an exam hall is "attached" to memories of walking to my seat, opening the door, arriving at the test center, etc. in an unbroken chain extending back to early childhood. Dreams, being episodes of varying length unconnected with one another, leave gaps in narrative that would identify them as dreams upon reflection. It may be maintained, in fact, that my ability to examine my memories and find the narrative chain is unbroken is indicative that I am not dreaming, at least as far as "dream" is normally defined.
It might be countered that in fact my entire life may have been a dream. Again, this remains a logical possibility that must be addressed, however the ability of my mental facilities to construct something as complex as the world which I experience seems highly questionable. Such a scenario is essentially a solipsistic one, in which my mind with no further input is constructing the world. This falls to the same criticisms used by Wittgenstein against solipsism; that is since I am using language to think (or dream) in, I must have interaction with other conscious agents in order to utilize correctly a rules based, essentially social system such as language. Dreaming being by definition a solo activity, the interactions with other agents would rule out the possibility of one's life being an extended dream.

Did anyone else get the e-mail inviting them to the new program? It had a link to confirm transfer, which took one to a trouble-shooting page...I got an e-mail back telling me that transfer would be avaible as soon as the software was up, and to reply to that message with any further questions; and of course my questions got kicked back as undeliverable...
I am a little confused about the registration, and I really hope that they get this straight very soon because I will be without internet acsess for a long time, starting in about three weeks...

Hey Plato, Casey, and all:

I received an email from London informing me that continuing registration was available online. There were two messages, actually, one giving me my old user name and a second giving me a temporary password, which happened to be my actual, old password. Uh, not the coolest thing to do on UoL's part, because, who knows, this is unencrypted text, and I might use that password for my online banking and trivial crap like that. Oh well. Anyway, I was able to get back onto the VLE, after agreeing to umpteen pages of terms of use, and changing to a new password. Everything's basically the same, what didn't work before still doesn't, and I wasn't able to discern how to register for next year. Thanks & hope your break from hard studies is doing well. --Ron