Tuesday, December 27, 2005

Surveillance: What Bush Should Have Done

Should all international communications be scanned by the NSA? Possibly, but not without warrants and accountability.

President Bush should have established automated mechanisms to scan messages for content, but he should also have created safeguards against use of the information by humans, and obtained the consent of Congress (which was in Republican hands, by the way).

What kinds of safeguards might be appropriate? Well, human translators might be permitted to analyze flagged messages, but they would only have access to the identities of the participants under court order. Also, the collected information should only be available for use in counterterrorism cases. Otherwise, this technology will be abused by people on fishing expeditions.

As it stands, Bush broke the law by illegally wiretapping communications. He wants us to trust him after all his lies. Trust him? Let's see, how did Bush put it? "Fool me once, shame on — shame on you. Fool me — you can't get fooled again."

I think there are even odds that Bush will be impeached if the Democrats take back Congress in 2006.

Thursday, December 22, 2005

Intelligent Design: The story so far

I've been debating ID online for the last couple of months. These are my conclusions so far.

The ID Claims
Some of the ID folks I have debated accept most of mainstream evolutionary science. Some even accept macroevolution, i.e., the idea that speciation is explained by natural selection and mutation. What they generally don't accept is that the pool of genes that is responsible for even the most primitive life could have evolved using undirected mechanistic processes. Instead, the sharper IDists focus on "front loading" - the idea that a designer programmed all the original genes, and mechanistic evolution took over from there. They're basically saying that a designer built all of the lego blocks, and natural processes built stuff out of those blocks.

Needless to say, they don't want to talk about the designer, who is presumed to be God. However, the term "supernatural explanation" is an oxymoron. The supernatural is a name we give to the unexplainable. Invoking the supernatural gives you a well-deserved one-way ticket out of the scientific enterprise. So ID proponents aim to find some scientific signature of design, but without reference to the nature of the designer.

To this end, IDists have tried to cook up some measures of whether or not a system is designed. The first measure is something called irreducible complexity (IC). A system is IC if it does not function if any one of its multiple parts is removed. For example, the blood-clotting mechanism in humans may fail if any one of certain proteins are not present. IDists argue that IC systems could not have evolved (at least not with high probability) because a system with multiple components could not have evolved incrementally, but would have to have evolved all at once.

As defined, IC systems do exist. However, the conclusion that they could not have evolved has been soundly refuted. For the ID claim to be valid, they have to show that none of the components of the IC system could have had any selective advantage outside the IC system itself. If a protein used in blood-clotting has no possible use outside the clotting mechanism, then ID has a more compelling case. However, if that protein has any other benefits to a life form, then evolution will have replicated the genes for making that protein anyway. It doesn't even matter whether life still has an alternative need for that protein. Though the protein may no longer be needed for anything except blood-clotting, as long as it was once useful for something, evolution has no trouble accounting for it.

The other idea that ID proponents refer to is something called complex specified information or CSI (yes, it's named after the TV show). CSI is very poorly defined. The concept has no credibility among mathematicians and scientists. ID advocate William Dembski has tried to cook up a formula for calculating CSI, but it has been solidly refuted. Dembski has claimed that information cannot be created through unguided mechanistic processes, but this is clearly false. Genetic computer algorithms do this every day by computing solutions to complicated problems.

Design as Science
Certainly, there are situations in which design is a scientific hypothesis. In archaeology and criminal investigations we regularly pursue design hypotheses. However, in these cases, we also know through direct experimental test that intelligent entities are present and available to do the designing.

What about cases where we have no other evidence of intelligence? The search for extraterrestrial intelligence (SETI) is a case in which we hope to detect intelligence in transmissions without otherwise observing the intelligent aliens. If we see a narrowband transmission containing a repeating sequence prime numbers, we're likely to think that it originated from an intelligent source. There are two things that would tip us off to the intelligence behind such a message. The first is artificiality - the signal is unlike anything predicted by alternative natural causes. The second is utility - narrowband transmissions containing mathematical data have utility to an intelligence.

Artificiality
Artificiality is sometimes easy to detect. If you find a intricate, mathematically precise structure made of a bulk material (like a titanium alloy), then this is a good signature of life, if not intelligence. Bulk materials like metals can easily be studied in the lab, and we can generally test the material to see that it doesn't spontaneously form such structures over a range of different formation conditions.

Yet, the artificiality of a structure is extremely difficult to detect when we don't have the knowledge or skillset to be able to synthesize a comparable structure in the lab. This is the case with the nanomachinery of living cells. Today, we have neither the knowledge nor the technology to fabricate an artificial amoeba. So, we don't understand the dynamics of cell structure formation or abiogenesis. For this reason, we cannot point to biological structures and say they are artificial because they are not expected under the laws of biochemistry. For all we know, such structures might have formed easily in Earth's primordial soup.

You can't claim the artificiality of a structure when you don't have detailed knowledge of the physics involved in its history.

Utility
Utility is a more compelling signal of design. Let's revisit our SETI example. Narrowband signals are harder to generate, but they use a lot less power, and power efficiency is useful. Also, a sequence of transmitted primes has utility as a universal greeting message for alien civilizations.

The same applies for the common examples of intelligent design by humans. For example, we can observe that a flint axe has utility to a creature that has limited intrinsic ability to cut flesh and bone.

Consider a simple steel girder. Such a girder has utility in construction. In fact, any collection of regularly shaped macroscopic objects will typically have a utility. We manufacture things in regular sizes so that we can apply rule-based procedures to them in their application or for commerce.

Language symbols also have utility for storing information. Even if we could not translate the text in a message, we might be able to perceive some utility to its design.

Now, evolution also predicts we will find structures that have utility. The one utility evolution predicts is survival. Evolutionary processes will generate structures that have selective advantages, i.e., that make their biological hosts fitter in their respective environments. So, if you're looking for signs of ID in biological life, you need to look for structures that have utility, but which have no survival benefit (neither for the species in question, nor for its ancestors).

So, if you look at a biological structure, e.g., the human eye, you cannot argue that its utility is enough to infer ID unless you can show that the human eye has no selective advantage (which, clearly it does have).

One problem for ID proponents who favor early-stage design (as opposed to ongoing intelligent tampering) is that sexual reproduction, mutation and natural selection will probably have pruned from ancestral life any feature which had no selective advantage. If an alien seeded Earth's life with programmed DNA, any part of that program that wasn't related to the system's survival would likely have been eliminated by natural evolutionary processes.

Prediction
When challenging ID proponents to make predictions, they commonly respond by saying that unguided evolution (UE) makes no predictions. IDists are making two mistakes when they do this. The first is in assuming that ID and UE are theories. They are not theories per se. They are classes of theories. A theory that aims to explain a facet of evolution using specific unguided mechanistic processes would be categorized as a UE theory. A theory that proposed that a biological system was designed for a particular utility by a particular designer using some specified technique would be a theory that falls into the ID category.

UE contains many predictive theories. UE theories predict inherited mutations, varying rates of speciation, patterns of adaptation, etc. UE theories are statistical. They cannot predict exactly what evolution will produce, only that it will produce something with statistical, selective advantage in its environment. UE is predictive because it contains theories that are predictive.

ID fails to be predictive because there are no scientific theories that fall into the ID category. In an effort to bypass the question of God, IDists ignore the question of utility. A discussion of utility would force them to speak about the utility of life beyond mere survival. Unless you know something about the designer, you cannot say what would be useful to him/her/it/them. This forces them to consider artificiality alone, and to hopelessly do so in a realm where the physics are poorly understood.

A Word About Probability
You'll hear ID proponents talk about the low probabilities that life formed by undirected mechanistic processes. Most of this handwaving is pure bantha poodu. I once saw a presentation in which the speaker spoke of the astronomical odds against a DNA molecule forming out of its constituent atoms in a vacuum! What does this have to do with evolution? Nothing. It just generates irrelevant numbers to throw at non-expert audiences. No one thinks that life formed by individual atoms flying together all at once. Atoms would have formed into precursors like amino acids, proteins, and membranes. More complex molecules would have formed from the precursors. Little research has been done in this area, so no one knows the odds of life forming in a planet-sized bath of such precursors. It could be unlikely, or it might be a virtual certainty. We don't know.

However, even if we did know that the formation of life is improbable under those conditions, one cannot use this as an argument for design. We are not independent observers of the creation of life on this planet. We are dependent observers. We are not free to observe life not forming on Earth 3.8 billion years ago. That is, the probability that life formed on Earth and led to our existence is unity.

This scenario is very different from us, say, observing that new life developed independently on another planet. In that case, our existence as observers does not depend on the formation of life on this other world. If we see life improbably popping up elsewhere, then we have a clue that we don't understand what's happening in abiogenesis.

This question of observer independence is closely related to the Doomsday Argument and the Adam and Eve Paradox. If you try to apply probability measures when you're not an independent observer, you run into all sorts of nasty paradoxes.

By the way, sound reasoning about probability reveals that there is no cosmological fine-tuning problem, either.

Conclusion
There are two scientific signatures for design: artificiality and utility.

Our ability to detect artificiality is limited by gaps in our knowledge of natural phenomena in biochemistry. Today, ID relies on these gaps in our knowledge of natural processes to suggest that artificiality lurks within biology.

The only utility we have observed in life so far is selective advantage. This is exactly what we would expect of life that was created through unguided evolutionary processes. As far as I can tell, there are no ID predictions about utility because God has no scientific utility for life.

No trace of ID has been found by either method. No experiment has been proposed to test an theory of ID because there are no actual theories of ID.

The frequent misbehavior of ID proponents cited by the Judge in the Dover case has raised the bar even higher for ID theories. ID theories (if we ever see any) will require extraordinary evidence in order to overcome the noise of all the false claims floating around in ID-land.

Where's the Outrage?

...asks Arlene Getz at Newsweek.com:
Tutu recalled teaching in Jacksonville, Fla., when Bush won re-election in 2004. "I was shocked," he said, "because I had naively believed all these many years that Americans genuinely believed in freedom of speech. [But I] discovered there that when you made an utterance that was remotely contrary to what the White House was saying, then they attacked you. For a South African the deja  vu was frightening. They behaved exactly the same way that used to happen here - vilifying those who are putting forward a slightly different view." Tutu made these comments to me exactly a year ago next week. I haven't seen any reaction from him about the latest eavesdropping revelations, but I doubt he is remotely surprised at the U.S. president's response: a defense of the tactic, together with a warning that the government would launch an investigation to find out who leaked the news to The New York Times.
For most of my life, I've been a trusting sort of chap. I really want to be able to give the government the benefit of the doubt when they say they'll spy on us with care.

However, Bush is asking us to trust him. And no one should. This is a man who has called the Constitution of the United States "a goddamned piece of paper." A man who has trashed international treaties and abandoned the Geneva conventions. He has explicitly championed torture. He's given the Pentagon the job of the State Department, and allowed the military to do domestic investigation of civilians.

Some will say that he's really pulling out the stops in an effort to protect America. They would be wrong. What about border control? Our southern border is totally porous. This makes no sense. His foolish attack on Iraq has weakened us militarily and diplomatically, and made terrorism far, far worse. Thanks to Bush, thousands of new terrorists have become trained veterans of urban warfare. The man is thoroughly incompetent, to say the least.

Ask yourself this question. Why does Bush need to illegally spy on Americans without a warrant when there is already a secret court designed to facilitate legitimate national security needs?
The 11-member Foreign Intelligence Surveillance Court is perhaps the most secretive in America, created by the Foreign Intelligence Surveillance Act (FISA). It reviews U.S. government requests for surreptitious eavesdropping to gather intelligence on potential U.S. enemies. Last year, according to a report to Congress, it received 1,758 warrant requests -- and approved all but four.
Bush has consistently tried to remove transparency and accountability from government. Here is yet another case where he asks Americans to trust him instead of trusting institutions.

Five years ago, the United States was the leader of the free world. The USA stood for freedom, human rights, technological innovation and liberal democracy. Now, we stand for spying, torture, imprisonment without due process, invasion without due cause, breaking treaties, fundamentalism, and incompetence. What's next? When do you say enough is enough? Is there any outrage that Americans will not tolerate?

Oh yeah. Consensual sex in the oval office.

Tuesday, December 20, 2005

I propose a toast!

To science!

The Intelligent Design movement suffered a major setback today. Not only did they lose the Dover, PA trial, they got a shellacking.

PZ Meyers over at Pharyngula has all the juicy details, including these quotes from the court's opinion:

Moreover, cross-examination revealed that Professor Behe's redefinition of the blood-clotting system was likely designed to avoid peer- reviewed scientific evidence that falsifies his argument, as it was not a scientifically warranted redefinition.
...
We therefore find that Professor Behe's claim for irreducible complexity has been refuted in peer-reviewed research papers and has been rejected by the scientific community at large.
ID is a politico-religious movement that aims to undermine the scientific enterprise by redefining it to incorporate the supernatural. That's basically what the judge said:
In his ruling, Jones said that while intelligent design, or ID, arguments “may be true, a proposition on which the court takes no position, ID is not science.” Among other things, he said intelligent design “violates the centuries-old ground rules of science by invoking and permitting supernatural causation”; it relies on “flawed and illogical” arguments; and its attacks on evolution “have been refuted by the scientific community.”
It's about time someone put a stop to their unpleasant scheme.

This evening, I'm going out to celebrate.

Wednesday, December 14, 2005

Inference and Bayes' Theorem

Say you want to calculate the probability that your theory T is correct given observations O. According to Bayes' Theorem, the probability of T given O is

P(T|O) = P(O|T) P(T) / P(O)

where P(O|T) is the probability of O given T, and P(T) is the prior probability of T without accounting for observations O. P(O) is the prior probability of O.

P(O|T) is the degree to which T predicts O. If T doesn't predict O then either (i) T predicts that O is unlikely/impossible, or (ii) T is uncorrelated with O.

Let's consider case (ii) where T is consistent with O but doesn't preferentially predict O.

If T is consistent but uncorrelated with O, then the probability of O given T is just the probability of O. That is P(O|T) = P(O). Consequently, P(T|O) = P(T). In this case, you cannot make any inference of T from O.

More intuitively, the probability that a theory T is inferred by observations O is proportional to the force with which T predicts O. If T doesn't preferentially predict O, you are not justified in making an inference from O to T.

For an Intelligent Design theory to be inferred from the data, it must be specific enough to predict something about the observations.

One objection to this conclusion would be to claim that an arbitrary T can always be fitted to O such that it preferentially predicts O. For example, one might claim that "T is the generic theory that there is an unknown agent that is responsible for O." However, the problem here is that T is no longer an inference from O. It is a paraphrasing of O. (After all, we originally set out to answer the question "what is the unknown agent responsible for O?")

So, how do we know whether or not T is just paraphrasing O? We can know this by counting parameters. If we have N data points, we need N parameters to paraphrase the data without making any inferences. For example, we can always write the next number in a sequence as the sum of the previous number and some parameter tuned to give us the correct answer.

Therefore, an inference is a theory T with fewer parameters in it than O. Since there are fewer parameters in T than in O, some proper subset of O must be predicted by T from the remainder of the data in O.

Generic ID makes no predictions, not even within the existing data we have. It has more free parameters than any amount of data we throw at it. It is, at best, a paraphrasing of the data.

We can certainly infer the action of intelligent agency from some data sets, but only when our intelligent agent theory is predictive in some way.

Monday, December 12, 2005

More on Inference and Explanation

Suppose my car won't start. Is there an acceptable explanation which explains how the car fails to start that does not also predict that the car will not start?

For example, I would predict that, if the fuel line were cut, the engine would not start. So, a cut fuel line sounds like an explanation, whether it is probable or not.

However, I cannot accept that my car will not start because "roses are red." Roses are indeed red, and their redness is perfectly consistent with my car not starting. However, the redness of roses does not predict that my car won't start.

Nor would I accept the "explanation" that an undetectable gremlin prevented the engine from starting. The undetectability of gremlins (and their unpredictable nature) renders the gremlin just another word for "ignorance of cause."

So, a plausible explanation is a picture of a cause and effect relationship. If the proposed cause is unknowable, or does not predict the effect, then you don't have an explanation.

Therefore, any valid explanation of a phenomenon must predict that phenomenon from knowable facts (i.e., from chemistry, physics, etc.).

To claim that there is no physical theory that explains a physical phenomenon is to claim that there is no explanation of that phenomenon at all.

Saturday, December 10, 2005

Pegasus Exists

I apologize for the length of the following brain dump. Writing is my way of forcing myself to think these things through.

Unlike The Golden Age of Balloning, there no fees or penalties imposed for not reading this post.

Radical Empiricism
I have really taken an interest in Radical Empiricism. I'm not buying into the philosophy of William James wholesale. Instead, I merely claim that thought and emotion are experienced, and should be accorded a more equal footing with the physical experiences of taste, touch, sight, sound and smell.

Suppose we redefine the term empirical to refer to any pattern which is experienced, i.e., not just patterns among the five senses. This radical empiricism asserts that all the things we think and feel are also empirical because we experience them. So, the very idea of, say, "Pegasus", commits us to its existence, at least among mental sensations.

Of course, in common use, we don't normally say that purely mental things "exist." So what is the connection between existence in this radical empiricism and existence in the everyday sense? It is clear that the more familiar use of the verb "to exist" signifies a correlation between a mental pattern and a physical one. Thus, in common use, claiming that "Pegasus exists" is the same as asserting that one can experience patterns in the five physical senses that correlate with the mental pattern of Pegasus.

The correlations between patterns of mental sensation and patterns of physical sensation need not be direct. Finding feathers with equine DNA would qualify as an indirect correlation between the mental and physical patterns that would be Pegasus.

Philosophical Existence
Philosophically, the question of existence gets more complex. The world of sensation appears to us "as if" there is an external world that impresses upon our senses.

In the scientific world, existence still retains the same meaning as it does in everyday use, namely, as a correlation between mental sensations and physical ones. This is why it is perfectly valid under radical empiricism to formulate a theory of quarks that explains the experimental data, even though quarks are not observable. Quarks themselves exist as mental patterns that correlate indirectly with physical observations.

Though we are permitted to ask whether mental patterns have corresponding physical patterns, we cannot ask whether there is a correlation between mental patterns and things that, even in principle, cannot imprint upon our senses. The question is simply malformed because there is nothing to correlate the thoughts with. Neither the verb "to exist" nor the noun "thing" has any meaning beyond experience or potential experience.

From this perspective, a multitude of metaphysical problems are just a result of misclassifying thought as something beyond experience, a classification which seems, at best, an artificial distinction among experienced things.

Consider platonism. Here's how the Stanford Encyclopedia of Philosophy briefly defines it:
Platonism is the view that there exist such things as abstract objects - where an abstract object is an object that is wholly non-spatial and non-temporal (i.e., that doesn't exist in space or time) and, hence, is entirely non-physical and non-mental.
Are platonists claiming the existence of things not experienced? If so, they would be taking language far beyond its realm of applicability.

Philosophical Inferences
How precisely can we state the limitations of knowledge given that our only window on the world is an empirical one? Historically, we have been very successful in identifying patterns in the world of physical sensation that are well-explained by structures that exist only in mental models (e.g., the aforementioned quarks). We have also been able to do the reverse. We have correlated physical observations with patterns of mental sensation using lie detectors or functional Magnetic Resonance Imaging.

What is not so obvious is whether it is possible to infer information from sensation that is wholly non-empirical. I claim that such inferences are impossible.

If I have a set of propositions P = {P1, P2, P3, ...}, I can only make an inference, N, from these propositions if N has some implications for the truth of my set of propositions P. That is, knowing N and some subset of P, I can derive the remainder of P.

Therefore, any inference I make from a set of sensations must have implications for that set of sensations. This rules out any possibility of inferring anything non-empirical from empirical facts.

Furthermore, when we say that a mental conception is "about something," we're saying that the conception correlates with some other sensation, mental or physical. This raises doubt about whether non-empirical propositions are really about anything at all.

What Doesn't Exist
The non-existence of a thing is the absence of a sensation that correlates with a mental conception. There are mental and physical things that don't exist. Prime numbers between 7 and 11 would be an example of a mental sensation that does not appear to exist. However, we can only say that we have searched for a thing and found nothing, not that there is no possibility that such a thing will be found. We can have very high confidence that certain patterns (especially some mathematical ones) will not be found to exist, but I have proposed no mechanism to guarantee certainty of non-existence.

Science, Physical and Mental
If there are well-defined patterns in sensation, what are the properties of such patterns, and how can we find them?

We perceive correlations in sensation. We can see a man on horseback, water under a bridge, water flowing downstream, and apples falling from trees. We can see the patterns of arithmetic in collections of objects, and the simultaneous presence of two sensations.

We can also correlate the correlations themselves. We can correlate the man on horseback riding across a bridge under which water flows downstream. What we don't see is a correlation of correlations that denies one of the correlated sensations. We don't simultaneously correlate the pattern of the man on horseback riding across a bridge under which water flows downstream with the pettern of the same horse with no rider.

There is a name for this latter kind of correlation: an inconsistent correlation. Not only do we perceive the world to be largely consistent, we have little use for inconsistent information. Any action that is inconsistently correlated with the experience of something desirable, might just as frequently produce the experience of something unpleasant. It would be fair to say that an action that has an inconsistent correlation with a particular outcome is not truly correlated with that outcome at all.

Logic is the procedure used for detecting inconsistencies in correlations. By correlating sensations with symbols, we can implement the rules of logic more formally as symbolic logic.

If there are useful correlations to be found in sensation, they must obey the rules of logic. The search for useful correlations is a search through all of the logically consistent mental patterns that can be correlated with observations. That is we are looking for a mental pattern that acts as a model of the world, so that our observations appear as if the world fits the model. This is why mathematics is important. Not only are mathematical structures empirical themselves, they also represent an enumeration of logically consistent models that can be correlated with a logically consistent physical world.

From here, it is easy to see why the scientific method works. We locate a logically consistent mathematical structure that can be correlated with our observations of the world, then we use that same structure to predict future observations and correlations. If the prediction turns out to be wrong, we modify the mathematical model and try again.

There is a very important empirical fact about mathematics that has implications for this analysis. Given a finite number of experiences and correlations, there are an infinite number of mathematical models that can be correlated with those experiences. If I provide you with some points on a graph, you can fit a curve to those points, but you cannot be certain that you have correctly fitted your curve to all the points that might appear on the graph in the future.

Meaning
What implications does radical empiricism have for meaning in language? Given the experience of a language proposition, we can only interpret it by correlating it with other experiences. This means that meaning in language is a scientific endeavor. No language proposition has infinite precision in its meaning because no correlation confers infinite precision. Instead, we form a scientific theory about the meaning of the proposition that has implications for how we expect to see the proposition used (correlated) in the future. That is, to have meaning, a proposition must be correlated reliably with something other than the symbolic sense data in which it is written. The meaningful proposition must claim definite implications, if it is true versus if it is false.

The extent of the implications of a proposition map out the domain of its meaning. If a proposition has implications only within a related set of propositions about mental sensations, then the proposition's meaning does not extend to physical patterns. The propositions of algebra do not, by themselves, have meaning that extends to the physical world. All propositions are evaluated within a specified context. This allows a proposition and its negation to be true in two different contexts without contradiction. The separation of contexts is what eliminates the simultaneous truth of the propositions from rendering the system inconsistent. Note that context boundaries may not be on sense boundaries. Two purely mental contexts may be totally independent, e.g., two algebra problems can exist without contradiction in their respective contexts, despite having axioms that would conflict if they were in the same context.

A scientific theory combines the propositions of mathematics with the empirical propositions of physical experience. The theory creates a new context in which the mathematical propositions have implications for the physical ones. This merged context is created by rules of correspondence between mental and physical sensations. In such a system, physical sensations are added to the system as new axioms. If the new empirical axioms render the system inconsistent, then the theory is disproved. If the new empirical axioms are consistent, then the theory is confirmed.

In contrast, a metaphysical system is one in which propositions about mental sensations are declared to have no implication for physical sensation. This is a denial that the metaphysical propositions share context with the physical world. However, creating this kind of separation is no different from the establishment of a mathematical context. Thus, at best, metaphysics is no different from mathematics, save for the evocative symbols it uses.

Friday, December 09, 2005

Inference

Consider the numeric sequence:

1, 2, 3, 5, 7, 11, ?, ?, ?,...

There are an infinite number of formulas that could generate this sequence.

Consider these conjectures:
1) The sequence is generated by an unknown formula that produces {1, 2, 3, 5, 7, 11} and follows it with indeterminate numbers.

2) The sequence is generated by a formula that produces {1, 2, 3, 5, 7, 11} and follows it with the sequence {6, 3, 2, 6,...}.

3) The sequence is an ascending list of integers evenly divisible only by themselves and 1.

The list of such consistent conjectures is infinite. Which ones do you infer?

I can see three criteria that we might apply to determine whether or not an inference is valid:

(A) A conjecture is inferred if it merely replicates the data, i.e., it is consistent with the data.

(B) A conjecture is inferred if it replicates some subset of the data, and makes a prediction about the future.

(C) A conjecture is inferred if it replicates some subset of the existing data from another subset of the existing data, i.e., it internally predicts the data.
If you go with criterion (A), then every data set should lead you to infer a solution like #1. But #1 isn't explanatory. You don't know anything more about the data (or the future) if you accept solution #1.

Solution #2 has the added virtue that it is predictive. However, it has more free parameters than the data set. You have no more reason to infer this theory instead of a different theory that will predict any other number as the next number in the series. This solution could be inferred by criterion (B).

Solution #3 meets criterion (C). You have reason to choose solution #3 based only on the data in the existing set. Any number in the sequence predicts the next and previous number in the sequence.

To summarize the criteria met by these solutions:




Criterion ACriterion BCriterion C
Solution 1YesNoNo
Solution 2YesYesNo
Solution 3YesYesYes


I claim:

i) that all three solutions are consistent, but not all are inferred, unless inference means the same thing as consistency. I reject criterion A.

ii) that a solution must be future-predictive to be scientific, but that future prediction alone is not adequate for inference. I reject criterion B.

iii) solutions that are internally predictive are inferred. Your inference should enable you to predict some subset of your data from another subset of your data. I accept criterion C.

This issue of inference came up in a recent discussion about Intelligent Design. My claim is that generic ID doesn't make any predictions, either internally or externally. Therefore, it's not even inferred, let alone scientific. However, if your ID theory is specific enough to allow you to predict, say, one aspect of the fossil record from another aspect, then you can make an inference. Your inference may be less than scientific, but it is, at least, an inference. To do this, you have to know enough about the physical limitations of the designer to say why the data is the way it is.

The next question is, are there any explanations that meet criterion B, but not C?

I'm guessing there aren't because any such solution wouldn't explain the existing data set. It would just be a wild guess about the future.

Tuesday, December 06, 2005

The 2004 Torture Referendum

A recent poll finds that sixty-one percent of Americans approve of torture. In 2004, I echoed the sentiment that the American people had a simple choice in the November presidential election. That choice was to vote for or against Abu Ghraib. At the time, I thought it a clever rhetorical device. It's now abundantly clear why this device failed to have much persuasive effect on my fellow citizens. They did vote on the torture issue, but in favor of George W. Bush.

Monday, December 05, 2005

Everything is Empirical

Yesterday, I found David Hume's book, An Enquiry Concerning Human Understanding at Project Gutenberg.

Published posthumously in 1777, the text remains lucid:
Every one will readily allow, that there is a considerable difference between the perceptions of the mind, when a man feels the pain of excessive heat, or the pleasure of moderate warmth, and when he afterwards recalls to his memory this sensation, or anticipates it by his imagination. These faculties may mimic or copy the perceptions of the senses; but they never can entirely reach the force and vivacity of the original sentiment. The utmost we say of them, even when they operate with greatest vigour, is, that they represent their object in so lively a manner, that we could almost say we feel or see it: But, except the mind be disordered by disease or madness, they never can arrive at such a pitch of vivacity, as to render these perceptions altogether undistinguishable. All the colours of poetry, however splendid, can never paint natural objects in such a manner as to make the description be taken for a real landskip. The most lively thought is still inferior to the dullest sensation. - Paragraph 11
It is common in everyday speech to separate the empirical (the things we experience in the world of the five senses), from the rational (the things that happen inside our minds). But isn't it better to consider every aspect of experience as constituting empirical fact?

If so, thought and imagination should also be considered empirical. We make discoveries in our minds by way of reason and explanation, and we remember our thoughts much the way we remember a day at the beach.

Breaking down the artificial separation between the rational and the empirical is liberating. As Hume explains above, our liberation costs us nothing in our ability to distinguish imaginary things from physical things.

However, we must surrender absolute certainty in making our escape:
There is a species of scepticism, antecedent to all study and philosophy, which is much inculcated by Des Cartes and others, as a sovereign preservative against error and precipitate judgement. It recommends an universal doubt, not only of all our former opinions and principles, but also of our very faculties; of whose veracity, say they, we must assure ourselves, by a chain of reasoning, deduced from some original principle, which cannot possibly be fallacious or deceitful. But neither is there any such original principle, which has a prerogative above others, that are self-evident and convincing: or if there were, could we advance a step beyond it, but by the use of those very faculties, of which we are supposed to be already diffident. The Cartesian doubt, therefore, were it ever possible to be attained by any human creature (as it plainly is not) would be entirely incurable; and no reasoning could ever bring us to a state of assurance and conviction upon any subject. - Paragraph 116
[Hmmm. Perhaps it is time to change my blog subtitle!]

Hume's skepticism does not deprive us of logic, mathematics, science or meaning. Hume's philosophy is infused with practicality. It merely admits that all conclusions we might reach in these endeavors have some degree of uncertainty. With diligence and repetition, the uncertainties of constructions like logic, mathematics and meaning can be made arbitrarily small.

Saturday, December 03, 2005

Historians: Bush Worst President Ever

An informal survey of historians suggests George W. Bush will be remembered as the worst president in U.S. history.

  • He has taken the country into an unwinnable war and alienated friend and foe alike in the process;

  • He is bankrupting the country with a combination of aggressive military spending and reduced taxation of the rich;

  • He has deliberately and dangerously attacked separation of church and state;

  • He has repeatedly "misled," to use a kind word, the American people on affairs domestic and foreign;

  • He has proved to be incompetent in affairs domestic (New Orleans) and foreign (Iraq and the battle against al-Qaida);

  • He has sacrificed American employment (including the toleration of pension and benefit elimination) to increase overall productivity;

  • He is ignorantly hostile to science and technological progress;

  • He has tolerated or ignored one of the republic's oldest problems, corporate cheating in supplying the military in wartime.