A quick list of some of the relevant issues and arguments we have considered. Most have been discussed in class; others are in the readings but have not been discussed much in class. This is just meant to be a starting point for your study. I don't pretend that this handout lists all the important arguments, and for those it does list, it certainly doesn't contain a full enough description to enable you to do well on the midterm!
 

What is it?

Intentionalism is the view that the qualitative characteristics of conscious mental states can be explained in terms of their representational properties.

Article

We read Michael Tye, "Visual Qualia and Visual Content Revisited."

Tye distinguishes between "qualia" and "Qualia" (447). "qualia" (lower-case 'q') are "the introspectively accessible properties of experiences that characterize what it is like to have them." Tye thinks that qualia in this sense can be explained as a certain type of representational property.

"Qualia" (upper-case 'Q'), by contrast, are "intrinsic, introspectively accessible, nonrepresentational qualities of experiences." Tye thinks there are no such things as Qualia in this sense. (The main source of difficulty is the term "nonrepresentational": Tye thinks that all the interesting properties of mental states are representational; defining "Qualia" as nonrepresentational rules out his favored theory of qualia by definition.

Tye briefly describes his positive view of qualia ("phenomenal content") in section III, using the helpful mnemonic PANIC: phenomenal content is "poised, abstract, nonconceptual, intentional (or representational) content" (454).

Arguments

If phenomenal character is phenomenal content, then it is impossible for two states to have the same phenomenal content but different phenomenal character (i.e. same phenomenal content but different qualia). Most of Tye's article consists of an examination of examples that seem to be examples of just this, followed by Tye's explanation of why he thinks the example is not a genuine counterexample.

Purported counterexamples, with Tye's replies:

experienceless sight: Tye argues that the experienceless visual information does not have the same content as that of a sighted person (less detailed).

inverted spectrum: Tye argues that even though a person with "normal" visual qualia may share all the same color judgments of someone with inverted qualia, there is still a difference in the content of their experiences: the normal person is perceiving things correctly and the inverted person isn't. (See pp. 451-452 for details.)

seeing two same-sized trees, one closer than the other: Peacocke suggests that we see both as being the same size, but they nevertheless look different. Tye responds that the "looking different" is still a difference in content: in addition to judgments about size, we also make judgments about "viewpoint-relative size." Despite having the same perceived objective size, one tree is perceived to have a larger viewpoint-relative size than the other.

Articles

Defense of externalism about content: Putnam, "The Meaning of 'Meaning'"

Explanation of narrow content: Brown, "Narrow Mental Content"

Arguments for Externalism

1. Twin Earth argument: I believe that water is wet. My Twin Earth doppelganger does not believe this. But all our intrinsic properties are the same, Therefore, what we believe is not determined by our intrinsic properties.

2. Argument from the linguistic division of labor: I don't need to know how to determine exactly what my words refer to as long as there are experts who can determine this. An individual's reference can be parasitic on that of the experts. But in that case the individual's intrinsic properties do not determine what he or she is referring to.

Arguments for Narrow Content

1. Causal argument: (1) mental states causally explain behavior by virtue of their content. (2) an entity's causal powers depend only on its intrinsic properties. (3) broad contents do not depend only on intrinsic properties. Therefore, (4) the kind of content that explains behavior is not broad content. There must be a different kind of content that depends only on intrinsic properties: this is narrow content.

2. Introspective access argument: (1) we have authoritative introspective access to the contents of our own thoughts. (2) we couldn't have such access to broad contents. Therefore, (3) our thoughts have narrow contents (and these are what we have introspective access to).

Argument

The book really advances one big argument, it seems to me, along with a lot of examples and supporting material.

The big argument goes something like this. Our use of "transparent" technologies to extend either our noncognitive or our cognitive abilities is no different in principle from our use of our built-in, neural equipment. The difference that matters is not (a) whether something is inside or outside the body, or even (b) whether it is part of our genetic endowment or something we have created. Rather, the difference that matters is that between transparent and opaque technologies, i.e. in the richness, bidirectionality, speed, and and reliability of our connections to the technology. As long as our use of certain "equipment" is transparent, it makes no difference to us whether it is built-in or added on.

If that is right, then we have two options. We can insist that every time we use any sort of equipment for moving or thinking (whether it is our own brains or external aids), we are using something external to our real selves. Or we can say that our selves literally include not only our internal equipment, but also other transparent technologies that we use.

The first option, while a theoretical possibility, would narrow the scope of the mind to the point that there wouldn't be much of anything left. If our minds consist only of what is actually present to consciousness, then most of the most interesting cognition we engage in is not actually done by "us." (If that's what we are, then epiphenomenalism begins to look like a plausible view.) So we should reject that possibility and accept instead the idea that we include whatever transparent technologies we use (and thus we are "natural-born cyborgs").

Theses

Dennett argues for a lot of conclusions in this book. Some of the main ones (I think): (1) Free will is not incompatible with either determinism or, more broadly, naturalism. (2) Even if indeterminism is true, it doesn't have any important connection with free will. (3) We can give a detailed account of how freedom could have evolved as a result of purely natural processes. (4) We should think of real freedom as arising from morality instead of the other way around.

Stages in the evolution of freedom. (Dennett's book is puzzling in various respects, and he doesn't explicitly lay out a specific series of stages like the following, so I'm probably missing some and misdescribing others. But he seems to have in mind something at least in the general ballpark of the following.)

1. "Choice" of biological characteristics. This isn't real choice, but a precursor. It is not engaged in by a single organism, but by a species over a very large number of generations. Should the species adopt or abandon a particular mutation? The species essentially tries both options out: one or a few organisms mutate; both they and the unmutated organisms pass on their characteristics to their offspring; the more successful version will produce more offspring and eventually the less successful variation dies out. Although this is not real choice, it shares some characteristics with real choice, for example the existence of multiple options and the existence of reasons to prefer one option to another.

2. Evitability. The evolution of things that can avoid other things (or destructive events generally). Clearly compatible with determinism and mechanism: e.g. a fairly simple robot could dodge missiles that were lobbed at it. Again, not real choice, but has some similarities with real choice: there are multiple possible futures (in the relevant sense); there are reasons to prefer one over others; an action of the organism causes the actual future to be one of the better ones.

3. Evolution of altruism. Depending on the details of a society, it may be to the advantage of individuals to act for the common good. (Not true in simple prisoners' dilemma situations, but more sophisticated models can explain the development of "altruistic" behavior.) One trick here: in general what is really advantageous to the individual is seeming good rather than actually being good: a successful freeloader will be better off than someone who genuinely does their part. The trick, though, is that under some conditions the most efficient way of seeming to act for the common good is to actually do so.

4. (Cultural) evolution of the ability to represent alternatives and reason about which one would be best. This becomes possible (at least to any significant degree) only with the development of language, which is a cultural artifact passed on by teaching rather than genetically. (Although of course it's only possible for an organism with the necessary biological equipment.) At this level the kind of process that species engage in by a kind of blind trial and error becomes conscious and internalized: we don't have to actually try out different options and see which one works best; we can simulate this process mentally and then just perform the action that turns out best in our mental simulation.

Turing suggests what has come to be called the "Turing Test" as a way of determining whether a machine is intelligent (or more precisely, as a better question with which to replace the ill-defined question of whether a machine is intelligent).

He then suggests that AI is possible, and responds to a variety of interesting objections to this thesis.

The Main Idea

Idealism is the idea that there is only one sort of substance, namely mental substance. There is no such thing as "material substance." Physical objects are collections of ideas (or, better, logical constructions out of ideas).

Hardly anyone defends this view in anything like this blunt form (but see John Foster, The Case for Idealism). Arguably, though, there are echoes of it in a variety of twentieth-century anti-realisms.

The Self-Contradiction Argument

We discussed one of Berkeley's arguments for idealism. Berkeley wants to show that an object cannot exist without being perceived. What he actually argues for is that an object cannot exist without being conceived. The argument is basically a reductio ad absurdum. You think there can be objects that aren't conceived? Give me an example! But your example of an unconceived object  is an object you've just conceived, so you're contradicting yourself.

I suggested that there were several things wrong with this argument.

1. Failure to distinguish between perception and conception. (Even if the argument showed that there couldn't be unconceived objects, it wouldn't show that there can't be unperceived objects.)

2. Failure to distinguish between thinking of a specific object vs. thinking of a property that many objects might have. (I can think "there are things no one has thought of" without thinking of any particular thing that no one has thought of. So even though trying to show that there can be unconceived things may require me to conceive them generically, it doesn't require me to think of them specifically.)

3. Contradiction in use vs. contradiction per se. A "contradiction in use" is a statement which I cannot utter truly because the fact that I am uttering it is inconsistent with its being true. Example (Armstrong): "No one is speaking now." But a proposition whose utterance is a contradiction in use need not be self-contradictory and therefore need not be impossible. It is entirely possible that no one is speaking now even though whenever I say it, it is false. (Similarly, "I do not exist" does not express an impossible state of affairs, even though I cannot say or think it truly.)

The relevance to Berkeley's argument is that, if he shows that the defender of unconceived objects is guilty of a contradiction, it is only a contradiction in use, not a contradiction in what he is describing.

4. Conceiving a fact vs. conceiving a thing. I can conceive that there are unconceived objects without there being an unconceived object that I conceive.

Berkeley seems to think that the defender of unconceived objects is committed to:

     There is an object such that it is unconceived and I conceive it.

That is indeed self-contradictory and therefore impossible. But the defender of unconceived objects does not need to be committed to it, but only to:

     I conceive that: there is an object that is unconceived.
 

Article

Paul Churchland, "The Puzzle of Consciousness" (chapter 8 of The Engine of Reason, the Seat of the Soul).

Argument

Churchland describes seven features of consciousness. He then describes a neural network in the brain, and suggests that it could account for all seven features. He does not argue that this is the correct model of consciousness, but rather that it shows how a neurobiological account could account for consciousness, and suggests that the correct account will probably bear at least some resemblance to the one he describes, although the details may differ.

The seven features are:

1. involves short-term memory
2. independent of sensory inputs (in the sense that we can remain conscious even when we aren't sensing anything, at least for a while)
3. displays steerable attention (what we consciously notice among our sensory data depends on what we're attending to)
4. capacity for alternative interpretations of complex or ambiguous data (e.g. images with more than one interpretation)
5. disappears in deep sleep
6. but is present ("reappears") in dreaming
7. different modalities of sensory experience are united in a single unified experience

Churchland suggests that a recurrent network can explain 1 and 2; 3 and 4 fit naturally with what we know from AI research about characteristics of neural nets; and 5-7 make sense given the neurobiological evidence about the way the network he describes functions.