One man's algorithm is another man's data.

Doesn't It All Just Come Down To Information?

What Even Is Information, Anyway?

"Information" is one of the great buzz words of the last several generations. The term has been in use in the English language for centuries, but it started to be used in its present technical sense in 1948, when a brilliant communications engineer working for the phone company named Claude Shannon published "A Mathematical Theory of Communication", ushering in the field of inquiry now known as Information Theory. He formalized the use of the term, and made it mathematically quantifiable. He thought of information as sequences of bits, or ones and zeros.

Claude Shannon was not a philosopher, he was an engineer. He mathematicized information so that he could calculate, for example, that a communications channel capable of transmitting X bits per second with an error rate of up to Y bits per 1000 could be used to transmit Z bits per second error-free (where Z is somewhat smaller than X), given some sort of transformation of the information on either end of the communications channel. He was concerned with noise on the wire. He was concerned with characterizing the "information density" in a given stream of bits so that by compressing the stream (i.e. increasing the information density) one could effectively transmit the same amount of information using fewer bits and therefore less bandwidth on the communications channel, thereby saving the phone company money. Essentially, Shannon was interested in very practical, meat-and-potatoes sorts of questions. Others, however, have not been so conservative. Information theory has inspired many philosophers to make extravagant claims, and information has become one of the most popular bases in the reductionist's toolkit. That is, just about everything at one time or another has been argued to be really just information, or information processing.

Of particular interest here, of course, are minds and consciousness. Indeed, the entire cognitive science program is predicated on the notion that the brain is (just) a complicated information processor - that not only can it be seen in terms of information processing, but that seeing it in these terms (or "higher-level" terms grounded in such terms) captures what is interesting about the brain in its entirety. A consequence is that any similarly configured information processor of equal capacity would manifest a mind in every sense that the brain itself manifests one. These sound like large and sweeping claims, but we can not even know whether they are or not (let alone whether or not they are true) until we nail down what is meant by "information" and "information processing".

In what sense does an information processor actually process information? How does it manipulate symbols? In spite of the well developed field of information theory, it is devilishly hard to find anyone who commits an actual definition of the term "information" to print. While qualophiles may not have answered the corresponding questions for the term "qualia", they acknowledge at least that there is work to be done along these lines. People on both sides of the Hard Problem debate, however, too easily assume that we know what we are talking about when we speak of information and information processing. Information is more difficult to pin down than is generally accepted, and there are very different things that are meant by the term depending on the context.

"Information" is a perfectly fine English word, and it has been in use for a long time. For all I know, Shakespeare may have used it. Everyone has a rough and ready, colloquial sense of what it means, and they use the word to communicate with each other every day. It also has this highly technical, bits-and-bytes-on-a-wire meaning. Mischief and confusion result from this mismatch, so if we are going to define anything in terms of information, we better be clear about what we mean, or at least we should have some distinct lines we can draw between information and not-information.

Information Is A Platonic Abstraction

There are molecules of ink on a page made of more molecules; there are perturbations in a physical electrical field on a metal wire; there are photons of light which propagate through an optic fiber. When I look inside a computer, I see voltage levels, and diodes which behave differently when subjected to different voltage levels. All of these things (or collections or patterns thereof) may be seen as information, but the key phrase there is "may be seen".

Information theory is a branch of mathematics, and bits (0s and 1s), like lines and points in Euclidean geometry, don't really exist, at least not out there in the real world. They are Platonic abstractions. We may profitably see things that are really there (like voltage levels) as information, and make generalizations, and hence predictions about those voltage levels based on our analysis, but the specific predictions we come up with will never be anything that we could not have, in principle have derived from a sufficiently detailed knowledge of the physical system alone without reference to any notion of "information".

Information is an abstraction, and abstractions, to a physicalist, must be cashed out in terms of the nuts and bolts that make up the actual physical universe. In practice, it may be very difficult for us to make useful predictions about an information processing system at the level of raw physics, but the universe itself has all it needs to clank along, one moment to the next, without our notions or theories of "information". Put differently, once God had established all the physical facts of the universe (i.e. the physical laws and initial conditions) He did not have to do any additional work to determine the facts about information processing. Everything the universe needed with regard to information was already baked in.

Information is always carried, or manifested, by something else. More pointedly, information always just is something else. By itself, information doesn't do anything. There is always something else doing the work, and that something else would do that work whether or not we think of it as informational. It is not merely the case that the information needs a substrate to instantiate it - the information just is the physical substrate, just as heat just is the mean kinetic energy of a collection of molecules. A system may be seen as informational, and we may thereby derive interesting and important conclusions, but these conclusions will themselves be may-be-seen-as conclusions, couched in terms of the abstractions of information theory.

But when people invoke the term "information" to describe some physical stuff interacting with other physical stuff, they are not usually talking about the stuff itself as such. Information is necessarily abstract. It is not the voltage levels or the ink, but the pattern of voltage levels or ink. As Rosenberg has pointed out (1998), the informational content of anything, whether ink on a page or electrical impulses on a wire, is a bare schema, or a pattern of bare differences. That is to say, the differences by virtue of which something is considered to be information are differences that are circularly defined in terms of each other. What is 0? It is not 1. What is 1? It is not 0. And this is all you ever need to know, all there is to know, about 0 and 1.

0 and 1 can be manifested, or carried, by any medium capable of assuming two distinguishable states (voltage levels on a wire, water pressures in a hydraulic system, wavelengths of light on an optic fiber). This substrate must have a nature of its own that outruns the simple criterion of distinguishability of states necessary to carry, represent, or manifest the abstract 0s and 1s of the purported information itself. One of information's distinguishing characteristics is that it is independent of its particular carrier. Information is arbitrarily transposable, or, to use a popular term, it is multiply realizable.

As I (and others) have argued, qualia are not arbitrarily transposable. Qualia are not themselves information, although they can carry information. Qualia are not a pattern of anything else, but the stuff of which patterns can be made, the substrate whose nature outruns the criterion of (mere) distinguishability. Redness is a qualitative essence and can not survive any transformation or translation into anything but redness. Some information could turn out to be conveyed by qualia, but qualia can't ever turn out to be (just) information.

Information Represents

Understanding, then, that when we speak of "information", we are not speaking about something real in itself, but rather as some good old physical thing that may be seen as carrying or manifesting information, what makes some physical things count as information and others not? It might be tempting at this point to turn from the strictly syntactic notions of information theory to a more semantic characterization of information. We might say that information represents something.

If we go there, however, we have left Claude Shannon behind. He and the phone company don't care what bits represent, or whether they represent anything at all. We are no longer in the quantitative, technical realm of bits, bytes, formulas, and information theory, and we have entered the squishier world of connotation, context, and intuition.

Back in the old days, I had an answering machine on my home telephone. When I didn't pick up the phone, it told whoever was calling that I wasn't home right now. It was a classic, purely causal, beer-can-falling-off-a-fence-post physical system. To what extent was it really, truly, representing me as not being home right now? How much more internal state would it have to have, "modeling" the world in some special way, perhaps processing this model in an "integrated" way, before we would say that yes, it really was representing me as not being home, in any way that was relevant to these discussions?

What does it mean for information to represent (without circular reference to information)? What do we mean when we use the term "represent"? What is the core intuition or experience that leads us to use the term the way we do? Does the light from distant stars, striking an earthly telescope, constitute information that represents the stars? Do all effects represent their causes, simply by virtue of the fact that someone might potentially be able to infer the cause (or something about the cause) just by observing the effect?

We might then start by saying that thing1 represents thing2 if thing1 is caused by thing2, or if thing1 varies in regular, lawlike ways as a function of variations in thing2. It is sometimes said that information is "a difference that makes a difference." But this is too broad to be any use at all. Since from the time of the Big Bang, each particle in the universe has some influence on every other particle (from the non-zero gravitational influence that any two objects of non-zero mass exert upon each other, if no other), everything is caught up in the causal mesh. Everything behaves just the way it does as a function of everything else (at least, everything else it its "light cone", if you want to be physically accurate). If information is anything which is caused by other things in lawlike, regular ways, then everything is information. In fact, everything is information about everything else. If everything is information about everything, then the term is nearly useless, and should be replaced, in philosophical debates, with the more honest term "stuff". And "information processing" could be reasonably be replaced with the expression "stuff doing stuff under the influence of other stuff".

Descriptive vs. Prescriptive Information

A great deal is made of the fact that information represents, but this descriptive, representative sense is only half of the informational story. There is a whole other aspect of information that plays a huge role in our lives and in our theories. Information comes in two flavors: 1) prescriptive ("pick that up.") and 2) descriptive ("the museum is open today"). Algorithms are prescriptive, data is descriptive. The algorithm operates on data. The opcodes that comprise a computer program at the lowest level are prescriptive information (they tell the CPU what to do during a given tick of the computer's internal clock), whereas the data upon which the program operates (whether that data comes from in the computer's memory or from outside, through an input device) constitutes descriptive information. Descriptive information represents (or misrepresents) something, while prescriptive information tells you to do something. If a fragment of a computer program says, "If x is greater than 43, open the pod bay doors", the fragment itself is prescriptive, while the number being examined, the x, is descriptive data. Those opcodes are purely causal, and themselves comprise absolutely everything a computer ever does. Their prescriptive nature is as blunt as that of a baseball hitting an antique vase. They just do.

In everyday conversation, we tend to think of information as primarily descriptive: it sits there, and you hold it before you and regard it: "Oh, so Bismarck is the capital of North Dakota. How interesting." But algorithms are information too ("Go three blocks, turn left at the light, pull into the Krispy Kreme drive-through and order a dozen hot glazed doughnuts."). As far as information theory is concerned, Shannon's laws, etc. don't care at all whether the information is taken as descriptive or prescriptive by the eventual receiver of the information. Any string of 0's and 1's has the same bandwidth requirements on the wire and is quantified exactly the same way whether regarded as descriptive or prescriptive, as data or algorithm.

If you find a computer file full of binary data, and you have no way of telling what the data was used for, you can not tell whether the file constitutes descriptive or prescriptive information. There is no fact of the matter, either, if you just consider the computer's disk itself as a physical or even an informational artifact. It's just a bunch of 1s and 0s. For you to make the prescriptive/descriptive distinction, you must know what the file was intended for, and in particular, you must know a lot about the system that was supposed to read it and make use of it. Only by taking the receiver of the information into account, and looking closely at how it processes the information, can we determine whether the file constitutes data or algorithm. Does the receiving system open the file and treat it as salary records, or does it load up the file and run it as a program? Indeed, one system could treat it as a program, and another could treat it as data, compressing it perhaps, and sending it as an attachment in an email message. The choice of whether a given piece of information is prescriptive or descriptive depends on how you look at it.

Example Using Boolean AND

Consider the AND gate. An AND gate is a very simple piece of circuitry in a computer, one of a computer's most basic logic components. It is a device that takes two bits in and produces one bit as output. In particular, it produces a 0 if either (or both) of its input bits is 0, and produces a 1 if and only if both input bits are 1. That is to say, it produces a 1 as output if and only if input1 AND input2 are 1. Note that the operation of the AND gate is symmetrical: it does not treat one input bit as different from the other: 1 AND 0 gives the same result (0) as 0 AND 1. Another way of saying this is that the AND operation obeys the commutative law. The operation of the AND gate is summarized in the following truth table:

input1input2input1 AND input2
000
010
100
111

But now let's arbitrarily designate input1 as the "control" bit and input2 as the "data" input. Note that when we "enable" the control input (i.e. we make it 1) the output of the whole AND gate is whatever the data input is. That is, as long as the control input is 1, the data input gets passed through the gate unchanged, and the AND gate is effectively transparent. If the data input is 0, then the AND gate produces a 0. If the data input is a 1, then the AND gate produces a 1.

When we "disable" the control input however, (i.e. we make it 0), the output of the whole AND gate is always 0, no matter what the data input is. By holding the control input 0, we turn off the transmission of the data bit. So the control input gets to decide whether to block the data input or let it though untouched. It is the gatekeeper. But (and here is the punchline) because of the symmetry of the AND gate, our choice of which input (input1 or input2) is the "control" and which is the "data" was completely arbitrary! The decision of which input is the prescriptive input telling the gate what to do with the descriptive input is purely a matter of perspective.

Information Pokes, Pushes, or Nudges

If we are speaking in terms of information theory, even loosely, we are in the realm not of conscious humans, but of systems. These systems can have an internal state, and they communicate using information, or signaling of some kind, traversing a communications channel. The communicating systems may exhibit behavior and/or change their own internal state based on information they receive or access. In this realm, the comfort zone of the computer scientist, the information theorist, and the physicalist, strictly speaking there is no such thing as representative, descriptive information - all information is ultimately prescriptive. Insofar as information has any effect on a receiving system or information processor at all (that is, insofar as it is informative), it makes the processor do something. The data in an MP3 is an algorithm that commands a machine to construct sound waves that make up the music.

Think of a given piece of the information as a physical thing, say a tiny area on the surface of a computer disk that is magnetized one way or another way, indicating a 0 or a 1. If this area is to constitute information at all, it must be causally efficacious. That is, something else must do something, or not do something, or do something differently, because of the particular way that area is magnetized. For the magnetized area on the surface of the disk to be informative at all, it must make something else do something, just as a rock I throw makes a beer can fall off a fence post.

This sounds pretty prescriptive. Nothing happens by virtue of information simply being itself. At some physical level, it always comes down to the information (or more precisely, the information's physical carrier or substrate) pushing something else around, forcing a change on some other physical thing. Moreover, any physical system that forced the same kind of state change on the part of the receiver would thereby constitute the exact same information as far as that receiver was concerned.

A computer does what it does because of an algorithm, or a program in its memory. This algorithm is prescriptive information. It consists of a series of commands (opcodes), and the computer does whatever the currently loaded command tells it to do. The computer itself (or its CPU) comprises the context in which the individual commands have meaning, or rather the background dispositions which determine what each command will make the computer do. The data that the algorithm processes may be considered descriptive information, but to the extent that the computer's internal state changes on the basis of the data it is processing, hasn't the data dictated the machine's state, and thus its behavior? "If x is greater than 43, open the pod bay doors": isn't x here an opcode, whose value tells the computer to open the pod bay doors or not? The "data" is either not there for you at all, or it makes you do something. It is the cue ball: it knocks into other balls and sets them on an inevitable course of motion. All data are opcodes.

The prescriptive aspect of the supposedly descriptive data in a computer is obscured by the fact that the data lacks a clear, stable context in which its effects are felt, whereas the same CPU tends to do the same thing each time when given the same opcode. The effects of different data are highly dependent on the current state of the machine. Nevertheless, after the data is read, the machine's state is different because of the specific value of the data, and the machine will behave differently as a result. The machine acts differently because of this data, just as it acts differently on the basis of different opcodes in its algorithm. There is no principled natural distinction between the information that comprises the algorithm and that which comprises the "data" on which the "algorithm" operates.

Self-Reference

If all information is, at heart, prescriptive, then what becomes of reference, or self-reference in particular? Lots of thinkers have been very interested in self-reference for the last century or so, but what is so special about it? Is it really so mind-blowing that I can look up "dictionary" in the dictionary, or that I can write "This sentence is false" on a post-it note? If information is prescriptive or algorithmic, then all supposed cases of referential loops turn out to be causal loops like the earth revolving around the sun, or the short computer program "start: do some stuff; go back to start".

A computer routine that is recursive is one that calls itself, like the factorial calculator. Recall that, for instance, 5 factorial (written 5!) is 5 × 4 × 3 × 2 × 1, or 120. The computer program to calculate that looks something like this:


factorial(input)   # Assume 'input' is a natural number!
{
    if (input is 1) then return 1
    else return (input * factorial(input - 1))
}

When called and handed a particular number as an input parameter, this routine calls itself with the next lower number, which also calls itself with the next lower number, then finally when the number reaches 1, it returns a 1, and the whole thing unwinds. This routine, then, is self-referential. But as far as the computer running it is concerned, there is nothing special or mind-bending about it. It neither knows nor cares that it is calling itself rather than a long series of separate routines. At each call, it just adjusts its Program Counter register to go wherever it is told to go, pushing some stuff on the stack. One hundred different routines, or one hundred calls of the same routine, it makes no difference to the computer. In this, the computer is right.

All Models Are Algorithms

There are theories of consciousness that regard consciousness as a product of the interaction of a system with an internal model within itself, kind of a homunculus without the infinite regress (See Metzinger 2003). (This sort of theory has a lot going for it. Roughly, when a control system (like our brains) evolves to a level of complexity where it includes a model of itself in its overall model of reality, any "self queries" are actually queries of this self-model. The fact that these queries and their answers bottom out in undecomposable primitive representations that seem ineffable is to be expected, and not mysterious, and certainly not a reason to go off the metaphysical deep end.)

But what sort of additional information does an internal model provide the larger system that it could not have derived on its own (given the external stimuli), and how does this additional information confer consciousness?

It seems that if we have a system that contains an internal model, we could optimize it a bit, and integrate the model a little more tightly into the rest of the system. Then maybe we could optimize a little more, and integrate a little more, all the while without losing any functionality. How would you know, looking at such a system, if it just didn't have an internal model anymore, or it did but its model was distributed throughout in such a way that it was impossible to disentangle it from the rest of the system? In the latter case, what power did the notion of the internal model ever have? The problems with thinking that there is something special about self-models are similar to those that plague Higher-Order Thought (HOT) theories: once you separate out some aspect or module as special to the system as a whole (whether you call that thing a self-model or a higher order thought) the specialness really comes from the communications channel between that module and the rest of the system, and we are right back where we started.

Internal Models As Black Boxes

Let us assume a conscious system that has a distinct model (either a model of itself, or a model of the world, or a model of the world including itself - whatever kind of model deemed necessary to confer consciousness). In good functionalist fashion, let us denote this in our schematic diagram of the whole system with a black box labeled "model". You ask it questions, and it gives you answers. Between the "model" box and the rest of the system is a bidirectional communication channel or interface of some kind. This kind of thing is often denoted in schematic diagrams as a fat double-ended arrow (like this: ⇔) connecting the "model" box and the box or boxes representing the rest of the system. Think of it as a cable, perhaps a very fat cable, capable of carrying as much information as you like. Let us call this interface, the cable itself and the conventions we adopt for communicating over it, the API (for Application Programming Interface, a term borrowed from computers). This API may be quite complex, perhaps astronomically so, but in principle all communication between the rest of the system and the "model" box can be characterized and specified: the kinds of queries the rest of the system asks the model and the kinds of responses the model gives, and the updates from external stimuli that get fed into the model.

People who believe in these sorts of theories generally claim that the rest of the system is conscious, not the model itself. Because, by hypothesis, all communication between the (purportedly conscious) rest of the system and the model takes place over the API, the consciousness of the rest of the system comes about by virtue of the particular sequence of signals that travel over the API. As long as the model faithfully keeps up its end of the conversation that takes place over the API, the (conscious) rest of the system does not know, can not know, and does not care, how the model is implemented. It is irrelevant to the rest of the system as a whole what language the model is written in, what kinds of data structures it uses, whether it is purely algorithmic with no data structures at all except for a single state variable, or even purely table-driven in a manner similar to Ned Block's Turing Test beater. It could well be completely canned, the computational equivalent of a prerecorded conversation played back. As far as the rest of the system is concerned, the model is a black box with an interface. Let us just think of it then, as an algorithm, a running program.

Once you separate the model from the rest of the system conceptually, you necessarily render it possible (in principle) to specify the interface (API) between the rest of the system and the model. And once you do that, there is nothing, absolutely nothing, that can happen in the rest of the system by virtue of anything happening in the model that does not manifest itself in the form of an explicit signal sent over the API. Anything that properly implements the model's side of the conversation over the API is exactly as good as anything else that does so as far as any property or process in the rest of the system is concerned. All that makes the model a model is the adherence to the specification of the API. The model is free, then, to deviate quite a bit from anything we might intuitively regard as a "model" of anything as long as it keeps up its side of the conversation, with absolutely no possible effect on the state of the rest of the system.

As any model-based system can be fairly characterized in this way, I have a hard time seeing what intuitive pull this class of theories has for its fans. Remember, what we are looking for is something along the lines of "blah blah blah, the model gets updated, blah blah blah, and therefore red looks red to us in exactly the way that it does." What magic signal or sequence of signals travels over that API to make the system as a whole conscious?

In information systems as traditionally conceived, there are no models, no representations, no data. It is all algorithm. As engineers, we may find it useful to draw a line with a purple crayon and call the stuff on the left side "data" and the stuff on the right side "algorithm" or "processor", but this is not a principled distinction. It is ad hoc, a may-be-seen-as distinction. Any theories of mind that depend on certain kinds of "models" or "representations" being operative then degenerate back into strict functionalism, since the models they speak of turn out to be just more algorithm, just as if they were utility subroutines.

The Algorithmic Intuition

Where does the intuitive appeal of philosophies like representationalism come from? Part of it, I think, is the idea that the system, the processor or algorithm, can respond dynamically to the representation, the data. We have a sense that the algorithm has a certain identity, and that to the extent that it opens the door and invites data in to manipulate its own internal state, it does so under its own control. This intuition loses some of its strength when you fold the "data" into the algorithm (hardcoding the data), however. If you take the data upon which the algorithm is presumed to operate dynamically and declare it to be just part of the whole algorithm, the algorithm doesn't seem quite so dynamic anymore.

Algorithms are deterministic. Or rather, their physical manifestations are exhaustively described by the laws of classical physics. They barrel along on steel rails of causality. If you look closely enough at them, there are no options open to them, no choices whatsoever. If I knock a beer can off a fence post with a rock, it falls to the ground. There is no way even of saying that an algorithm runs correctly or incorrectly. There is no sense in saying that an algorithm is true or false. It neither represents nor does it misrepresent. It just does. (Or rather, and importantly, whoever or whatever faithfully executes the algorithm, plus the algorithm itself, just does. The algorithm itself just sits there).

The intuition that there is a certain plasticity inherent in algorithms, that they could do other things than what they do is a mirage. If I don't throw the rock, the beer can will stay on the fence post. While it may seem that an algorithm could behave differently given different data to operate on (if x equals 23, the pod bay doors stay closed), it would also behave differently if some of its subroutines were rewritten (if x equals 86, activate the espresso maker).

When people speak of algorithms manipulating representations, and look to them for the special sauce of consciousness, or anything philosophically big and fundamental, they are projecting intuitions about the mind outward into other stuff. Outside of certain limited technical contexts, the whole idea of the algorithm is a sneaky modern form of animism, an attempt to breathe life into cold dead Shannon information, made of Newtonian physics, to make it jump up and run around, to give it some inherent motive power, while denying motive power to the "data".

The Data Intuition

And what of our intuitions about data, as opposed to the algorithms that process it? Let's introspect for a moment and ask what is going on in our minds when we are aware of "raw" information, stuff that we imagine just sits there for us to know or perceive. (Without going down the rabbit hole of distinguishing between (mere) belief and knowledge, I am going to use "knowledge" here as a generally appropriate term that captures what I mean by the somewhat clunky "descriptive information" in our minds.) What is that like, and what makes us think we possess "descriptive" information or knowledge?

Starting with a qualitatively loaded example, I know that fire is hot, even when I am not near a fire. What does this particular piece of information consist of in my mind? How would I describe it? I would probably say something like, if I reached my hand out into a fire, it would burn me; if I put a piece of metal in a fire, it would get hot, and it would burn me.

I also have a piece of information in my mind that Paris is the capital of France. If asked to describe this knowledge, I would say that if I got on an airplane and went to Paris, I would end up in the capital of France, and I would have a whole lot of experiences that validated that. I know the cast resin garden Buddha is hard, and I know this with certainty - this is a piece of information I possess. What does it mean that I know this? I have an immediate, palpable sense that if I were to touch it, if I were to drum my fingernails on it, if I were to rap it with my knuckles, it would feel hard.

There is a common theme here, and that theme is hypotheticals. There is a whole lot of "if, if, if" in my descriptions of my own knowledge. Some of our expectations regarding these hypotheticals are immediate and sensual, while others are complicated and a little more abstract. I know that I have a certain balance in my checking account: if I tried to buy a roller coaster for my back yard, the debit would be declined.

If/then clauses have a decidedly algorithmic, prescriptive ring. One associates them with computer programs. To resolve them, you run through the cases. You compute. Could it be, in fact, that we do not actually know in the direct sense that we think we do, for instance, that the garden Buddha is hard? We only cognitively judge ourselves to know, and have a very good system for coming up with justifications on demand? If this were true, our "knowledge" of something is really just a warm, fuzzy confidence that we know rather than what we normally think of as true, immediate, internalized descriptive information. As I pointed out above, in the antiseptic realm of information processors communicating over a channel or accessing internal memory, there is no such thing as truly descriptive information, if "descriptive information" only ever actually informs by being activated and poking, pushing, and nudging. In the case of our minds, could it be that this activation involves running the "descriptive information" through some imaginary cases and getting results?

When does complete, just-in-time predictive power and confidence of your mastery of the hypotheticals become essence? How do you know that you know, really? Even something as seemingly definitional as 2 + 2 = 4? You feel certain that you grasp the meaning, and its inherent truth, all at once, but this is an appeal to introspective intuition. As a qualophile, I'm all for appeals to introspective intuition in such cases, but qualophobes often engage in intuition-shaming, so we should, in fairness, subject this one to the same sort of scrutiny.

This take on descriptive information is analogous to what Daniel Dennett thinks about qualia. He claims that we don't actually directly experience in the way we think we do, but we (merely) judge ourselves to experience. We actually have a really good mechanism for answering any questions immediately about our field of "experience", and we tell ourselves cognitively that we experience "directly". Could our own knowledge of things be that way?

What Is It Like To Know?

No, and for the same reasons that Dennett is wrong about qualia. I can know that the Buddha is hard, and really sense my knowledge of its hypothetical hardness without actually taking the time to run through any of the imaginary scenarios of touching, drumming, rapping. I've already talked about how odd it is that we can have a single thought that has temporal extension or flow built right into it, and how a smeared-out process becomes a unitary thing, grasped all-at-once. In our minds, the prescriptive becomes descriptive. Process becomes thing. We see the algorithm from above, without running through the if…then… cases like mice in a maze. For us, if…then… is not a matter of execution paths, but a more holistic, from-above ifthenishness. The counterfactuals are not just our way of expressing or explaining our knowledge, but are right there, baked into the knowledge itself, and into our sense of having that knowledge. There is a what-it-is-like to know the Buddha statue is hard. I know the Buddha statue is hard with the same sort of certainty that I know that it is hard when I am actually stubbing my toe on it. I am directly acquainted with my knowledge of its hardness as a piece of descriptive information.

Moreover, I know that I know it is hard. The troublesome second-orderliness of knowledge mirrors that of qualia: seeing red seems inseparable from knowing that you see red, just as knowing that Paris is the capital of France seems inseparable from knowing that you know that Paris is the capital of France.

In fact, (and if you have been following along you probably saw this coming), I'll take it to its next logical step: knowledge is a quale. Like a lot of qualia, it is a complex all-at-once kind of quale. Interestingly, it is also a Lego-stackable quale, in that it constrains or modifies, or calls into being, other qualia. Knowledge applies itself on the fly as the situation calls for it, or seems to present an opening for such application, and incorporates all those implied hypothetical scenarios instantaneously in some way, so that they don't actually have to play out through time in your mind. The ways in which a piece of knowledge can construct or constrain other thoughts you might (or might not) come up with is an inherent part of the knowledge itself. Pieces of knowledge seem to insert themselves and stack and self-organize as appropriate. They are active, and interactive.

Descriptive Information Is A Projection, And Weird

The intuition that an algorithm stands aloof, and regards inert data and makes choices based on it but not dictated by it, and that data and algorithm are somehow different, is an anthropomorphism. We project our own subjective, introspective experience outward. It's not wrong or silly! It only becomes silly when we try to bleach out any trace of the source material. We do create and use descriptive representations in our minds. We experience this every moment. We, as conscious minds, have a strong sense of having a separate identity from the simulations of reality we create and tinker with in our heads. We feel that we stand back from our models, our data, our pieces of descriptive information, and regard them, and make decisions based on them.

This sense, however, isn't quite as trustworthy as it seems. As William James said, the thoughts are the thinkers. As I have discussed elsewhere, the apparent distinction between the self thinking and perceiving, and the stuff thought or perceived, must be something of an illusion, under pain of infinite regress of the homunculus in the Cartesian Theater. Whatever the self is, it must incorporate any "descriptive" information it is aware of into itself as part of itself even as it thinks of itself as standing back and regarding thoughts, percepts, and bits of knowledge. The players are the audience, after all. It is this illusion, this fantasy image, that we project onto algorithms and data.

If physicalists want to deny qualia as fundamental, they should examine information too. They must give up the algorithmic intuition (doing vs. representing): that certain information does stuff and has any choice about what it does, and that certain other information doesn't do anything but is done to. To a physicalist, all information is purely prescriptive, deterministically so. Which is fine, but "information" then becomes either weak or (philosophically) boring, and it becomes pretty hard to say that consciousness all comes down to information (or the processing thereof).

There are descriptions in the universe. They just aren't information, in the strict, Claude Shannon, Information Theory sense. That is to say, information takes on its descriptive, representative aspect only when we create it and take it in all-at-once; when in our minds, it is something other than a cluster of dispositions manifested in a series of hypothesized scenarios to be played out algorithmically, and is, rather, a single thing, a partless whole. Importantly, a single partless whole that in some funny, as-yet poorly understood way, incorporates those dispositions and scenarios in a qualitative all-at-once comprehension. This ability of ours, as I have argued, is a unique, spooky mysterious thing minds and only minds do, like seeing red. As with the redness of red, it is hard even to talk about it in precise terms, which is all the more reason to try to talk about it, being honest with ourselves about the limitations of our usual ways of talking.