John Horgan has an excellent new blog that he has recently started up, called The Scientific Curmudgeon. Horgan may be best known for his provocative 1996 book The End of Science, which was one of the first books for the general public that expressed skepticism about string theory (another was David Lindley’s 1993 The End of Physics). His portrayal of Witten in the book was a bit of an unfair hit job, but he got the story of what was going on in particle theory about right, unlike just about every other science writer working at the time. In recent years his attention has turned to issues of neurobiology and cognitive science, as well as the relation between science, religion and mysticism. He now teaches at Stevens Institute of Technology in Hoboken, and runs its Center for Science Writings.
Horgan describes himself as a “hopeful skeptic”, writing:
I still see science as our best hope for understanding ourselves and the universe, and for creating, if not a sci-fi utopia, then at least a much better world. Scientists can provide us with cleaner, cheaper sources of energy; better treatments for cancer, AIDS and other diseases; more detailed accounts of how brains make minds. That’s why, in spite of writing a book called The End of Science, I’ve remained in the science-journalism racket, why I work at a science-oriented school, why I encourage young people to become scientists. But I also encourage greater recognition of science’s limitations and fallibility. It is precisely because science is so consequential that we must treat its pronouncements skeptically, carefully distinguishing the genuine from the spurious.
One of his recent postings discusses the issue of the Templeton Foundation, yesterday’s is a charming story about his daughter, linked with a tale of his adventures among the cosmologists back in 1990.
Is “hopeful skeptic” a reasonable position on String Theory, as opposed to, say, “hopeless skeptic”? I like Horgan’s science writing, and blog, but am not clear why he is hopeful.
Some things never change… 😉
A (possibly off-topic) comment posted on http://motls.blogspot.com/2006/06/reviews-of-certain-book.html#links and disappeared after 10 seconds.
hi, I am LubosMotl. At line AB of page XZ Woit put a “,” while Witten would have used a “;”. This proofs that Woit is a moronic crackpot. He makes me vomit. I predict that when I will be back, the censorship committee will have deleted my post.
Dear banned by Lubos who is LubosMotl,
what an intelligent comment.
Thanks for moving back to the right blog where you belong 😉
dear Lubos, thanks for saving western civilization from crackpots.
I love sarcasm.
Peter, have you actually met Lubos in person ? I wonder what would happen if you two ran into each other. You’d probably cave in to the years of societal conditioning you’ve been subjected to and crank up those fake smiles, greet each other politely, have a short, awkward chat, all the while pretending you actually like and respect each other. I guess that’s the civilized thing to do, as defined by societal norms.
Society sucks 🙁
I wonder what would happen if you two ran into each other.
– Modified from Douglas Adams, “The Hitchiker’s Guide to the galaxy”
While reading Horgan’s writings I had a thought about why public should be sceptical (if not suspicious) about theoretical sciences, especially when the experimental part is stagnant. It seems to me that scientific process has a huge amount of inertia built into it. There is a very low threshold for canonizing theories and a very low tolerance for alternatives. I don’t think this changed much during last 2000+ years.
We all laugh at medieval astronomers who couldn’t accept the heliocentric system. Aren’t we laughing at ourselves? According to present day standards the Ptolemy’s theory was a very respectable scientific achievement, worth defending against such “crackpots” as Copernicus and Galilei. It had a beautiful math behind it. There was a lot of experimental support. It was written in all textbooks. It was supported by smartest people (Aristotle) during millenia (not just 30 years). What else could you ask for to canonize a theory? How is it different from modern canonical HEP theories (I am talking not just about superstrings)? When will we ever learn?
runge_kutta and Christine,
Yes I have briefly met Lubos in person, and it was a perfectly pleasant encounter. It’s also a little-known fact that I once wrote a recommendation for a fellowship for him. But back then he was just an over-zealous fanatic (my recommendation said something like: “I don’t think you should support string theory, but since you’re going to do it anyway, you might as well fund a smart true-believer”), he hadn’t yet been driven over the edge into nuttiness by the failure of string theory.
Your writings are a lesson to a young woman blogger.
Louise (A Babe in the Universe)
Woit, I have been reading the back and forth between you and Motl for the past couple of months. I have found it quite interesting to deconstruct what is at the heart of you arguments. I have concluded that the heart of the problem is Horgan claims science is versus Motl’s viewpoint. That is what divides the two of you is a total disagreement about what constitutes fundamental science. While I am not qualified to discuss string theory, I have thought a fair amount about what constitutes good and bad science. Below are some comments about the scientific method that are at variance of what was found in “The End of Science”. I have somewhat a contrarian viewpoint, so have included somewhat extensive comments from a paper I am revising. If you think these comments add some clarity to you monologue, please post them, if not don’t feel you have insulted me. I wish you both well in you monologue about science, I have found this foodfight too distracting and need to return to other matters:
Scientific theories have three important characteristics: compression, usefulness, and generality. Without compression, there is no advantage to developing a theory, tabular results or taxonomies would be sufficient. Taxonomic sciences such as biology and paleontology are still scientific, A technical discipline that has large amounts of data but little interpretation contains little information since there is no surprise associated with new data. information gain cannot be measured by acquiring new data. Usefulness has two components: engineering application and predictability. Without the ability to apply a theory to accomplish calculations, the theory becomes a tautology without an exterior world to which it applies. It becomes mathematics. Predictability is a component of usefulness, since predictions lead to the ability to design rather than just build. Most theories advanced provide compression and usefulness. We also require theories to have generality. Falsifiability plays a key role in probing the generality of theories (Popper 1959). The importance of falsifiability is the role it plays in the understanding and development of proper science has been often misinterpreted by commentators rather than practitioners of the scientific method. Falsifiability should be properly understood as the complement to generality. While it is often claimed that a theory can be disproved by a single negative observation, this is not the correct interpretation. A negative observation is simply defines limitations on the generality of a theory. In other words, falsifiability does not negate theories, but rather it limits the boundaries within which theories are applicable. Since usefulness is an important criteria, there is always a boundary associated with the theory in terms of applicability. There is, of course, a point where repeatedly falsifications cause one to reject a theory (Jaynes 1968).
A single positive observation is much less compelling test than a negative one, since almost any single prediction can be verified by appropriate selection of evidence. There is an intimate connection between observation and sampling in the sense of Shannon’s formulation of communication theory that is needed for all sciences. A principle of ontology is advanced by espousing a specific type of deconstruction of uncertainty (Sardis 1995) and the nature of uncertainty is always an important component for understanding ones discipline. In general, the more information gain we receive from data, the more biased (in the colloquial sense) we should become because informative priors converge in a parametric sense to the correct result (Jaynes 2003). Thus, a theory of sampling needs to be extended when we go beyond the reduction to numbers and instead accomplish a reduction to symbols. Most sciences would benefit from a deeper exploration between sampling and confirmation that a symbolic theory of sampling relative to their particular subject would accomplish.
There are several aspects of confirmation that should be clarified relative to the symbolic. If when measurement reduces to a number, the process is symbolic in the sense that is not usually considered from both an experimental and computation viewpoint. All computation and measurement is interval based. One does not measure a point, one measures an interval number associated with the interval of experimental uncertainty. The same thing occurs when one is engaging in computation as was pointed out first by Young (Young 1932). Thus confirmation is less related to theory since the instruments used for confirmation and computation within the theory are not formulated within the theory. Another issue is the compression. The ontology associated with algorithmic information theory (Chaitin 1987) suggests that a theory that doesn’t have redundancy built into is not a good theory. But redundancy is not so easily discussed beyond the realm of the rational numbers. What redundancy means when one discusses abstract symbols is less clear. Issues such as interference and correlation are always possible between symbols.
Theory plays a different role than experiment; it writes sentences that in some cases explore well beyond the restrictions of instrumentality. Theory, while driven by the words and grammar of experiment, is always trying to introduce new forms, establish metaphors, and use words in less experimentally important context, so theory is more concerned with trying to ferret out meaning (mostly by detailed working out of specific examples or models). It does this by concerning itself with constant refinement of the explanations for what is known, establishing metaphors that connect previously unrelated subjects, improved tool making (both new mathematical concepts that aid theory and instruments that change what it is possible to measure), and making predictions. Theory, while keeping in mind Feynman’s dictum, allows one to define and explore semantics of the subject while being ground in the syntax imposed by experiment. Mathematics provides the theoretician with an entirely syntactic language to explore and maintain rigor. But as has been noted before syntax is not science, metaphors are how we establish the “truth” and metaphors are inherently transformational or substitutions by their nature. Thus, syntax may tell us whether a transformation is correct but semantics is necessary to decide if the transformation results in something meaningful.
Truth is primarily established through experiment. But truth also has a logic associated with it. Philosophers have tended to ignore such concepts as metaphors and tool making in their critique of science, and giving short shrift to the concept of refinements based on new methods for accomplish measurement that revolutionize the ability to observe (Dyson 1999). All measurements are metaphors of a sort, though many treat measurement as a tablet from Moses. Measurement is canonized by theoreticians in their writings, yet those who have dealt with data are much more circumspect about it. At some point, axioms relative to what and how we measure will be ferreted out by mathematicians. Once this has been accomplished, then we will have to face some logical limitations. Already we know that there are problems in physics that are undecidable in the sense of Godel (Wolfram 1985). But the question is whether there are more fundamental limitations than that. If the axioms which encompass the experimental methodology are sufficiently rich to encompass arithmetic, then there are experimentally undecidable issues in physics and in science. This is troubling–“Is science decidable?” in an experimental sense is far more troubling than a specific theoretical result that is undecidable. Equally unsettling is question of complexity that was first raised by Godel in the theoretical sense (Godel 1936). He noted that there were problems that had proofs that were intrinsically long, so long in fact that some theorems could never be established within a given set of axioms. He further noted that by adapting additional axioms that proof’s length might be shorted considerably. What if conformation by experiment has similar difficulties? What good is a theory that can never be confirmed by instruments because the number of measurements that would lead to its acceptance is longer than human civilization’s life span? Could this not be the case when and if we ever find a theory that has the possibility of being a “unified theory of physics”? I wonder if this message from the Landscape, if it becomes more real. If we adopt speedup axioms, then what are the consequences for our epistemology? The symbolic, from a experimental viewpoint, is fraught with potentially unpleasant consequences from a logical and complexity perspective.
If the symbolic viewpoint was more universal, then most of what constitutes science and its boundaries would have to be reexamined. This is why the ongoing monologue between Woit and Motl is so rich in terms of its implications about what is and isn’t science. I find that science is its infancy in terms of understanding what constitutes science because of the issues of information gain, conformation, sampling, as well as the theoretical aspects of metaphor. Drawing on Feynman’s “It is surprising that people do not believe that there is imagination in science. It is a very interesting kind of imagination, unlike that of the artist. The great difficulty is in trying to imagine something that you have never seen, that is consistent in every detail with what has already been seen, and that is different from what has been thought of; furthermore, it must be definite and not a vague proposition. That is indeed difficult.” Imagination in science is like writting a sonnet, but richer and more difficult.
The rich possibilities of at what point an intellectual discipline that can be confirmed or falsified by data suggests that concepts and boundaries of science have to be rethought. Sampling and confirmation based on data or the lack thereof need to be completely rethought from the symbolic perspective. The algorithmic perspective needs to be incorporated as well, but not necessarily the way Chaitin has suggested. In many ways, we are closer to a beginning of science and determination of “The scientific method” rather than “A scientific method” as currently exists. Understanding why science is successful from a ontological and metaphysical perspective, remains to be accomplished. If I had any doubts about us being at the beginning of understanding of what science is, they have been erased by the strange and amusing postings I have seen here. Best wishes to you all and thanks for all the fishes.
I like John Horgan thinks that widely accepted assumptions, like inflationary theory, are crap, and he has good reason to say that if you simply do all of the demanded science, rather than science that’s strictly limited to the belief that it requires a naked singularity to produce a big bang.
Projecting backwards to the point where it requires inflation to account for the horizon problem should be a clue that there is a mechanism that enables a universe with volume to have a big bang. Especially since the problems of causality and structure that are associated with the horizon and flatness problems don’t exist when a universe that has volume has a big bang.
I continually notice that people keep calling for accepted assumptions to be, once again, reassessed, but eveyone has their own ideas on which assumptions are open to review, because, for example, the flawed theory does do some things quite well. What they apparently fail to recognize is that the solution to fixing a fundamental flaw is in the foundation below, but not in some more grand theory that exists beyond the the flaw.
I think it was Urs who fairly recently commented in the research group to someone that, “this isn’t 1917″…
Yes, it is.
“this isn’t 1917″… Yes, it is.
Great point! I think this is precisely where the boundary between science and theology lies. As soon as we declare a certain physicist or certain theory beyond criticism, we stop doing science and switch to theology.
I like this quote from Oppenheimer (which for an untrained ear may sound as a definition of crackpotism):
There must be no barriers for freedom of inquiry. There is no place for dogma in science. The scientist is free, and must be free to ask any question, to doubt any asssertion, to seek for any evidence, to correct any errors.
As for me, I don’t mind to discover that we are still in 1905.
And in his case, free to vaporize, fry and otherwise irradiate a lot of people.
There is irony in John E. Gray writing:
“Drawing on Feynman’s “It is surprising that people do not believe that there is imagination in science. It is a very interesting kind of imagination, unlike that of the artist. The great difficulty is in trying to imagine something that you have never seen, that is consistent in every detail with what has already been seen, and that is different from what has been thought of; furthermore, it must be definite and not a vague proposition. That is indeed difficult.” Imagination in science is like writting a sonnet, but richer and more difficult.”
That’s because Feynman did in fact coauthor a sonnet with me, which was published by AAAS, republished, and set to music. My Physics professor wife and I, for that matter, had a poem in the magazine Science. Hence mainstream science journal referees and editors have openly shown ability to recognize scientific imagination in the methodology of poetry. For that matter, NASA commissioned me to write a poem as the frontispiece of a conference proceedings. Gray is exactly right about imagination in this context. The issue in this blog becomes something like: “Do String Theorists have too much imagination, in inventing pretty math that defines 10^500 universes in a multiverse; or do they have too little imagination in not being able to follow Feynman’s dictum ‘ it must be definite and not a vague proposition?'”
Or is the problem that String Theory is not Science, but rather Science Fiction or Poetry or Theatre, but does not meet Woit’s criterias for GOOD Science Fiction or Poetry or Theatre? Are we waiting for a happy ending, a lovely final rhymed couplet, a final act that wraps up all the subplots? Or are we Waiting for Godot?
Pingback: The Scientific Curmudgeon » Blog Archive » Pulling the Plug on Strings