"Once again, finding an anatomical location for processes that are crucial in the control of what happens next is one thing, and locating the Boss is another; anyone who goes hunting for the frontal display screen where the Boss keeps track of the projects he is controlling is on a wild goose chase." -- Daniel Dennett, _Consciousness Explained_
That quote is pretty characteristic of the whole book. Note that the first half seems fairly reasonable and precise, while the other is full of messy metaphors. This deceptive mixture of precision and "hand-waving" is pretty much the whole argument, and I guess he hasn't been upping his game since then, either.
Dennett's work is largely focused on ignoring, dismissing, and disproving the existence of the Cartesian Theater in any form whatsoever, but many intellectual disciplines depend on the Cartesian theater, including mnemonic techniques such as that of Giordano Bruno and visualization and other "induced hallucination" techniques as used in the fields of problem-solving P:D and creativity. While he claims to not be using the "anesthesia" of the Behaviorists, this seems a glaring omission in the phenomena he presumes to explain in full.
Having read through Dennett's work and made a few notes, I have to wonder why he bothered to write it, except for the need to pay the rent. Dennett is arguing for Strong AI (minds can be replicated as software in classical IC computers), as distinct from a Strong Mind position (quantum or other Dualism), such as that of Sir Roger Penrose. Isn't it too early in the game to start staking out positions when there is (still!) essentially no formal, material, scientific evidence for either position? But getting back to Dennett, we see him offering explanations of how the human mind works, and why it must be able to be replicated in ordinary computing technology. Let's review, shall we?
The different stages of the development the human mind (Chapter 7):
- Maintain homeostasis and the boundaries between the organism and the environment (Self-preservation)
- Anticipate events implicitly in the processes of metabolism (Plants perform optimization computations for gas exchange and utilization of starches in anticipation of night, computations implicit in the balance of organic components of the plant. True fact.)
- Explore the environment. (Novelty and broad forms of attention.)
- Plasticity of the brain and basic, hard-wired instinctive responses (Adaptation)
- Development and learning (Internal and external discovery of conditions, focused attention)
- Software (Data nutated by hard-wiring into software algorithms)
- Performance and "Spectation" (Observation as much for the purposes of learning and optimization as for control "in the moment")
- Auto-programming (Internalize and Identify principles as self-image / self-component)
- Algorithm and meme design from abstract principles ("top-down")
- Meta-algorithms (Scientific and Philosophical methods of inquiry and argumentation; mnemonics; memory castles; Cartesian Theaters)
As soon as we get to point 6 above, we start to see where Dennett's own concepts begin to undermine his arguments against even the existence of Qualia and the Cartesian Theater. Neither qualia nor the Cartesian Theater need to be a hard-wired part of the brain, but need only exist as (pick one: memes, algorithms, ideas, software). Need I mention that the whole of the book and all its arguments are just such memes? Every time Dennett asks the reader to see, hear, smell, taste, or touch a stimulus that isn't physically present (which he does, quite a few times), he is implicitly making use of your mastery of a Cartesian Theater, including a wide variety of qualia, while explicitly denying their existence the whole time! Once we accept that software of some kind exists in the brain/mind system, why can't we have something along the lines of a full software environment, like a Java Virtual Machine, a virtual brain within a physical brain?
Dennett's conceptualization of computing is rather misleading as well, relying on the readers' presumed ignorance of the subject to baffle with bullshit rather than dazzle with brilliance. Take a gander at these little gems!
First, for those of you who don't know, once upon a time, there were analog computers! They were prone to all sorts of problems that the digital, discrete-value computers we use today avoid easily, such as asymptotic functions. Think of all the error messages you get when you try to divide by zero. Even a calculator gives an error message, and full programs have error handlers that either disallow zero as an input for some variables, or neatly dodge dividing by zero in the midst of processing data. These are exactly the sorts of problems we see in the human mind/brain system in terms of runaway processes, from schizophrenia to epilepsy. Perhaps my examples are poorly chosen, but the point is that the brain is more likely to use analog, continuous values as opposed to digital, discrete values. Dennett fails to even mention this little reality, as it weakens the fantasy of life immortal, uploaded into a massive digital computer.
Second, he fails to mention that many special processes used to diagnose programming errors allow the display of the software's step-by-step functioning. BASIC programs have the TRACE function, and I'm sure that someone more computer-savvy than myself could mention such a function for nearly any programming language, as it is too handy a tool to do without willingly. Such functions give a summary of the high-level commands without getting too deep into the guts of the machine, in a way very similar to our own inquiries into our thought-processes when we are baffled by how we made that one silly mistake.
Third, Dennett fails to mention how operating system "shells" allow for a variety of access levels, and for processes to run simultaneously, with perhaps some in the background, and some in the foreground. So yes, you can have several things on your mind at the same time, and when you allocate much more attention to one process, you run the risk of having others not perform as well as they should. You can also practice self-deception, by spawning a function ("persuading others") with a lower level of access to the facts.
I could shine a light on one dark corner of the problem after another, but you get the idea. Dennett guiding us through the forest of computing technology is either disingenuous or dangerously overconfident.
Much later in the book, Dennett actually raises some interesting (and damaging) points. Late in Chapter 9, section 3, we're finally introduced to a concept, a "global workspace", that looks suspiciously like a "Central Meaner", or at least a Cartesian Theater, but it isn't central, its distributed throughout the brain cortex. This is where Dennett's earlier efforts to smear several different concepts together begins to become a liability to his argument rather than an asset. Only a handful of less popular and extreme views of Mind require a Central Meaner, but a Cartesian Theater makes the human mind special as well, and here Dennett is practically handing it to us on a silver platter!
So just what is this "global workspace" that Dennett finally admits to? Considering that something like 90% of the cells in the brain are basically a mystery to us, in terms of their function, that could be just about any (or several!) kinds of cells, and any mechanism at all -- including quantum. That's pretty much the end of the argument there for Dennett; if any of these cells do something that computer circuits and software cannot do, do it too well, or do it too fast to be automated properly, then consciousness can't be uploaded into a conventional computer. If there are a great many quantum processors, and their efficiency is high enough, then it will be a long time before even a quantum processor, as we understand it today, will be able to match the human mind.
Argument 1: There is a formulation of Determinism which actually outlines the extent and limits of free will and consciousness, rather than prohibiting them, such as, say, the juxtaposition of many potentialities, with little "hard", deterministic influence from historic precedents or trends. The mapping of the options can be deterministic, but the options are deterministically balanced. The more extensive the mapping of knowledge and options to be exercised, the "freer" the will and consciousness become, something I would hope that most observers would verify from their own experience.
Argument 2: The subject/object duality does not apply to either free will or consciousness, as each is a relationship of subject and object; a process rather than a state. (4 or more imaginary numbers for state, and an equal or greater number for abstractive processes, and the possibility of the transposition of values, giving many geometries for one unvarying mathematical object; choices become choices of deterministically equivalent geometries)
Argument 3: Consciousness and Free Will are forms of speculative processing, and involve data-structures which are deterministic, pseudo-random or truly random ((random numbers in computing)), and such data-structures are self-nutating, giving rise to surprising transformations of historic premises and trends: this "surprise" is Free Will, and Consciousness is the capacity for Free Will. This is one of the reasons why we assign most animals a limited capacity for Consciousness; we are not often or deeply surprised by their behavior.
Here are the parts of the book I found (surprisingly) helpful:
In Chapter 4, Dennett's concept of "Heterophenomenology" is extremely useful, makes for a great demonstration of the power of the Cartesian Theater once we admit its existence, and demonstrates many of the concepts I wish to make clear in terms of the Phenomenal World. If I am indebted to Mr. Dennett, it is here, as elsewhere, that he makes it clear that I believe in (nearly) everything he does not.
Chapter 7, the evolution of consciousness is also worthwhile, without too many caveats.
The rest of it I don't have much use for.