Why don’t climate scientists use Feynman diagrams?

This paper, purporting to prove that the atmospheric greenhouse effect violates the second law of thermodynamics and therefore cannot exist, was apparently published in the International Journal of Modern Physics — not the world’s leading journal in the field, but hardly a crackpot publication.

Among the gems in the paper (spotted here) is this (pp 59-60 of the arXiv version): Climatological radiation balance diagrams…

1. cannot represent radiation intensities, the most natural interpretation of the arrows depicted in Figure 23, as already explained in Section 2.1.2 and Section 2.1.5;
2. cannot represent sourceless fluxes, i.e. a divergence free vector fields in three dimensions, since a vanishing three-dimensional divergence still allows that a portion of the field goes sidewards;
3. do not fit in the framework of Feynman diagrams, which represent mathematical expressions clearly defined in quantum field theory [159].
4. do not fit in the standard language of system theory or system engineering [160].

I kid you not. So radiation balance diagrams should fit in the framework of Feynman diagrams or system engineering diagrams?

In case you’re wondering: section 2.1.2 supplies some basic definitions of radiation intensity and flux. Section 2.1.5 says “In classical radiation theory radiation is not described by a vector field assigning to every space point a corresponding vector…” but in fact mentions that the “modern” (not so modern, actually) description uses the Poynting vector, and also talks of black body radiation and its variation from real spectra. Nowhere do I see any relevance to point 1 here. As for point 2, so what?

Also, in case you are wondering, reference 159 is to the standard text on quantum field theory by Itzykson and Zuber, and reference 160 is to the SysML site. Apparently climate scientists should be specifying their Feynman diagrams in SysML.

More here (that blog takes the credit for spotting the Feynman Diagram thing: I didn’t wade through the first 60 pages of it myself!)

The authors apparently submitted this paper in 2007 and it got accepted in 2009. It reads like a prank (after Sokal, I suppose one sees prankery everywhere). But apparently they are serious, and apparently the journal editors didn’t care to take a second look at it, despite its bombastic title.

A new international science centre

I am attending the inaugural conference of the International Centre for Theoretical Sciences, (ICTS) [update 29/12: fixed expansion], a new centre of the Tata Institute of Fundamental Research. Though the centre has been in virtual existence since 2007 and has already organised several meetings, its foundation stone was unveiled today by C N R Rao. Prior to that, the centre’s director, Spenta Wadia, described the history of the centre and its goals. The evening concluded with a public lecture by David Gross on the nature of theory, the second of three public lectures in the conference.

What is the ICTS? It aims to be, primarily, a facilitator of collaboration and interaction among scientists — theoretical and experimental, from India and from abroad — and people from other walks of life, too. To that end, it will assist in organising conferences and meetings (and, eventually, be a venue for those meetings); host visiting scientists, and their students and postdocs, for extended stays; and also have a small core faculty of its own, as well as adjunct faculty from elsewhere. Three existing institutes were specifically named as role models: the Abdus Salam International Centre for Theoretical Physics in Trieste, Italy; the Newton Institute in Cambridge, UK; and the Kavli Institute of Theoretical Physics in Santa Barbara, CA, USA. The goal is to rival those centres in international impact, and Spenta Wadia and the others behind ICTS can certainly do it, with help from the Indian academic community and the government. The Karnataka government has donated 17 acres of land to the institute, near the picturesque village of Hesareghatta (which is also known for Protima Gauri’s Nrityagram dance centre). It will be exciting in future years to have an international centre of the calibre of ICTP, KITP or the Newton Institute in our own backyard (if Bangaloreans will excuse this Chennai-ite for that phrase).

Key to this, I think, is implementing what Prof Wadia and others talked about: the importance of diversity in interactions. From that perspective, I find the inaugural conference disappointing already. The conference is titled “Science without boundaries”, but out of 20 scheduled talks (not counting the public lectures), 11 are physics, 6 are mathematics, one is computer science, one is physics-dominated chemistry, and one is synthetic biology. [Update 29/11: One physics talk today was replaced by a computational biology talk, because of the snowstorms in the US.] Only 5 of these talks are over, and a few of them were excellent and I am sure some of the remainder will be excellent too, but I find it an extraordinary distribution of topics for a conference that claims to be on “science without boundaries”. What makes it worse is that the three non-physics non-math talks, and the sole experimental physics talk, were all crammed into today’s session, so that the remaining 3 days will be devoted entirely to theoretical physics and mathematics. Other than a nod to computer science, there is no theoretical talk in the entire schedule that is not physics or mathematics! Meanwhile, of the 11 physics talks, no fewer than four are on string theory — an interesting theory no doubt, but an entirely unverified one to date — and only three are on condensed-matter, surely the most exciting (and interdisciplinary) area of physics in recent years. Statistical mechanics, non-linear dynamics, and other areas of physics are entirely unrepresented. If the goal was to demonstrate the diversity and interdisciplinarity of “theoretical sciences”, a better attempt could have been made.

(On a related note: of the 11 India-based invited speakers, 6 are from TIFR and affiliated institutes. India has more diversity than that in the “theoretical sciences”.)

Be that as it may, the talks so far have largely been good, as have the two public lectures that have taken place (the third is scheduled for Wednesday, December 30). Yesterday Avi Wigderson gave an interesting talk on computational complexity theory, and though I think he was guilty of needless oversimplification as well as needless controversial statements in places, it was a good introduction to the question of whether P=NP. In his defence, also, it should be noted that he was a last-minute replacement for Sir Michael Atiyah, who could not make it because of illness; and he abandoned his holiday in Hampi to deliver this talk.

Today David Gross talked on the nature of theory, and it was an entertaining history of theoretical science, as well as a well-argued presentation of the importance and relevance of theoretical work. He made the point, emphatically, that revolutions in science don’t sweep aside previous knowledge, but build on it. (So Newton’s theory of gravitation is still useful, even though not quite correct.) Among other interesting points, he raised the possibility that even if a “theory of everything” (a unified theory of the fundamental forces, including gravitation) exists, it may be too sophisticated for human minds to grasp (just as quantum mechanics is too sophisticated for dogs to grasp). But he made it clear that he does not believe that.

I cannot resist quoting a mild dig Prof Gross made at the S in the name ICTS. He related the story of a press conference at Princeton after John Nash won the Nobel. Supposedly, a journalist asked Nash: “You have won the Memorial Nobel Prize in Economic Science. Is economics a science?” He replied: “No, any discipline that requires the word “science” to be added to its name is not one.” Social scientists, political scientists and creation scientists, take note.

Wednesday’s talk is by Albert Libchaber, on the origin of life. If you’re in Bangalore, do attend. Details here.

Purism

A few days ago, the eminent vocalist T M Krishna wrote an article in The Hindu bemoaning the increasing use in Carnatic music of instruments that are not sufficiently sensitive or flexible to reproduce the fine microtonal modulations (“gamakas”, similar to “meends” in Hindustani music) required to fully express a raga.

This is not a new controversy. The western “equally tempered” chromatic scale is an approximation that gets every interval except the octave subtly wrong. For example, the fifth note (“panchama”, “pa”) is supposed to have exactly one and a half times the frequency of the tonic or “sa”, but in the chromatic scale it turns out slightly flat: 2^(7/12) = 1.498 approximately. It has been recognised since Pythagoras (and even earlier, probably) that notes in small-integer frequency ratios sound pleasant when played together, but all intervals on the piano, except the octave, are irrational (being powers of the twelfth root of two). The compromise is necessary in western music to enable modulation to new keys without retuning the instrument. In Indian music, where the tonic is never changed, it is argued that the chromatic scale is unnecessary. While the piano is rarely used, the harmonium has achieved considerable popularity (it continues to be widely used in Hindustani music, though not in Carnatic music) but is detested by some purists for this reason.

Equally important, instruments like the piano cannot “bend” the note: they play a fixed pitch when you press the key. In Indian music, it is common to “slide” from one note to the other and “waver” about certain notes: it is such ornamentation (the “gamakas” and “meends” that I referred to above) that gives ragas their individual character.

So Krishna’s tirade against the keyboard and the saxophone is understandable but not new. (He also acknowledges that some imports of Western instruments have been successful, in particular the use of the violin since the 19th century, and the use of the electric mandolin by U. Srinivas.) Nevertheless, I wonder if an excess of such “purism” may not be detrimental to the music itself.

Most of the arguments against keyboard instruments can be, and have been, levelled against the santoor and its first and foremost practitioner in Hindustani music, Shivkumar Sharma. Shivkumar cannot produce genuine “meends” but he achieves the illusion of doing so with rapid sliding tremolos. He also produces several new dimensions to his music by striking notes simultaneously with both hands: his left-hand patterns, while his right hand plays the melody, are a form of harmonic accompaniment to the music — but done so tastefully that only the rigidest of purists would object.

Krishna names no names but, since he believes no experiment involving the saxophone has worked, he clearly has a poor opinion of Kadri Gopalnath, the best-known saxophone player in Carnatic music. Personally I am not fond of Gopalnath’s music myself, without being able to quite define why. But I found the contrast of the following two passages interesting. Krishna:

The saxophone, among other instruments, is today very popular. The artists have made some modifications and changes to try and make it sound Carnatic but that has just not happened. The inherent limitation of the instrument makes the artiste limit his choice of ragas. Is this necessary? This is, to me, ridiculous. Are we willing to limit the bandwidth of a musical idiom to accommodate an instrument?

Gary Giddins, writing on a recent collaboration between Gopalnath and American jazz saxophonist Rudresh Mahanthappa:
Gopalnath, who generally plays in a yogalike seated position, has perfected something that jazz saxophonists have been attempting for decades: moving beyond the Western chromatic scale into the realm of microtones, a feat harder for wind instruments, whose keys are in fixed positions, than for strings or voice. Jazz players, such as Ornette Coleman, John Coltrane, and Albert Ayler, had gone about it by varying intonation, blowing multiphonics (two or more notes at the same time), or squawking in the upper register, where pitches are imprecisely defined. Gopalnath does none of that. Using alternate fingerings and innovative embouchure techniques, he maintains faultless intonation while sliding in and out of the chromatic scale.

Krishna is of course much more knowledgeable on Carnatic music than Giddins, or Mahanthappa, but I wonder if he is not too close to the picture to see it fully.

There are few significant practitioners of the piano in Indian music, but one recent pianist who has attracted much recent attention is Anil Srinivasan, Krishna’s cousin. Anil is a trained western classical pianist who has been immersed in and absorbed Carnatic music since childhood. He has been grappling with the hard (many would say, impossible) problem of combining Western harmonies and Carnatic melodies, and — in collaboration with the vocalist Sikkil Gurucharan, in particular — has produced some very interesting music. I have no idea whether Krishna was referring to Anil in his article. But Anil is trying to do something very different from Shivkumar Sharma or Kadri Gopalnath. He is not trying to play authentic Indian classical music on an apparently unsuitable instrument. He is trying to use his instrument and his knowledge of harmonic theory to supplement, and expand, the music itself, while relying on collaborators like Gurucharan to supply the “purist” component of the music. He is playing to the strengths of the piano, not to its weaknesses. He is not trying to do things on the piano that are easy on Indian instruments: he is doing things that would be impossible on Indian instruments. Whether those things are desirable is a matter of opinion, but I — and many others, it seems — like the result.

It seems that Anil wrote a rejoinder to Krishna, which I haven’t seen and can’t find online (update 27/12: here), but Krishna’s reply accuses Anil of missing the point of Krishna’s criticism. The point, stated as narrowly as possible, is that playing Carnatic music on instruments incapable of expressing all nuances of the music, and therefore limiting one’s repertoire to “suitable” ragas, diminishes the music rather than enhancing it. Again, I’m not sure whether he is applying this criticism to Anil, who is not (I think) trying to reproduce the intricacies of the music, but to supplement and expand it.

But is Krishna’s narrowly-stated criticism justified at all? I can’t help thinking of another form of music I am fond of, the blues, which permeates jazz and rock as well as being an art form in its own right. The blues is defined by the “blues scale” and its “blue notes”: some notes — notably the third and seventh, and sometimes also the fifth — are played slightly “flat”, but not quite as flat as their Western “minor scale” counterparts. That sort of flattening cannot be reproduced on the piano. “Bending” and “wobbling” those notes is also integral to blues, and certainly cannot be achieved on the piano (it is achieved on the guitar by bending the string with the left hand). Yet nobody would argue that the piano has no place in blues or jazz. Since the early 20th century pianists have worked around their instrument’s limitations in innovative ways, and their experiments have enhanced the music, not diminished it.

Of the artists mentioned above, I usually like listening to Shivkumar, I don’t much enjoy Kadri Gopalnath, and I find Anil Srinivasan’s approach very promising. This is not a technically informed evaluation of these artists’ respective merits. But I think musical quality is not necessarily correlated with technical purity. There are lots of musicians who follow all the rules but are mind-numbingly dull to listen to.

It seems to me, on reading Krishna’s first article closely, that he is more worried about the effect of keyboard instruments on music education — but then there are much graver problems to worry about. A scholarly article on how music should be taught would be welcome. But complaints about unsuitable instruments sound, to me, more reactionary than artistic. T M Krishna is a well-established and widely respected musician, but he is too young to sound like a curmudgeon.

RIP, Asheem

Asheem Chakravarty of the band Indian Ocean died yesterday of a heart attack. Obituary here (and elsewhere).

Though not formally trained, he and his bandmates defined a new sound that combined elements of Indian folk and classical with Western rock. Asheem played tabla, and sang, often simultaneously. The combination of his tabla, the acoustic guitar of Sushmit Sen, the bass of Rahul Ram and drums of Amit Kilam combined to unique effect. In the early days, the band played largely instrumental music, but as time went on, the somewhat classical-sounding vocal of Asheem and the somewhat more raucous, folk-inspired singing of Rahul Ram dominated their music more and more. After struggling with college-festival gigs for nearly 10 years, the band achieved considerable success with their album Kandisa in 2000, and in recent years have been touring heavily around the world.

Asheem had a heart attack in October, in Doha airport on the way back from a tour of the US. He had been in a poor condition since then. While Indian Ocean have continued to honour their recent concert commitments with fill-in players, it is hard to imagine them without Asheem. I hope the band finds a way to go on. And my condolences to Asheem’s family, and to the band: though he was taken away early, he enriched many lives.

More thoughts on Apple

We’ve now had the Mac Mini for over a month. It’s mostly my wife using it, and finding it very easy to use. But here are some random observations:

  • Did I hate proprietary software? I thought I did, but I suppose I only hated the Windows world. There is no open-source program that compares to iPhoto, for example. (And do we really care if such programs are not open-source? I increasingly realise that I’m for open standards and interoperability, not necessarily open source as such.)

  • Is Mac OS X stable and reliable? Yes. No crashes yet. Perhaps recent versions of Windows are equally good, I don’t know.

  • Is the command line still useful? Yes. Example: my wife had over a hundred files in the “Downloads” directory and wanted to move only the JPG photos to the “Photos” directory on the desktop. Pointing and clicking to select them in the graphical file manager was a tedious process. I simply opened a terminal and typed “mv Downloads/*.jpg Desktop/Photos”.

  • But is the graphical interface that hides all the Unix complexity good at what it does? Mostly, yes, but there are slip-ups. Example: the default shortcuts for switching “spaces” (“virtual desktops” on Linux, no equivalent on Windows) were shown as “^ left” and “^ right”, and despite some Linux experience, my wife didn’t immediately realise that “^” was shorthand for the “Ctrl” key. (I suppose only power users want virtual desktops anyway, but she knew them from Linux and found them convenient.)

  • Is open source software still useful on the Mac? Yes. Case in point: VLC, the media player. Unlike Windows (last I checked), the Mac plays DVDs out-of-the-box — but enforces a region code, which may be changed only five times before it is permanently locked. This is strictly a software restriction, at the behest of the movie industry. VLC disregards the region code so we can happily play DVDs from multiple regions. We have also installed The Gimp and Inkscape: the alternatives were Adobe Photoshop and Adobe Illustrator, but those are expensive and the free programs are good enough.

  • Anything we found we can’t do, or not easily? Yes: save a flash video from YouTube into a separate file. (I suppose there would be plugins for Firefox to do it, but first we’d have to install Firefox.) On linux, look in the /tmp directory for a file whose name begins with “Flash”, and copy it. Supposedly it used to be a similar mechanism (different location) in earlier versions of OS X, but on Snow Leopard we couldn’t find it.

  • Do I plan to trade in my Linux laptop for a Mac? No.

One white car

‘Twas tee, enn, ten, vee, five, seven, six, two,
One torrential white car in driving rain.
It hit her, but it well could have been you.

A call-taxi halted to let her through.
Her umbrella flapped. And, bull-like insane,
Came Tee, enn, ten, vee, five, seven, six, two.

Muddy puddles trap one’s wet feet like glue.
The white car halted, as she tried in vain
To step past them. It well could have been you.

“Hm, why should pesky pedestrians do
Road-crossings in my car’s own service-lane?”
thought Tee, enn, ten, vee, five, seven, six, two.

He revved, and bumped her knee (he missed her shoe).
“Don’t cross so slow,” he said; sped off again.
It’s his road. Does your knee belong to you?

No injuries? Not even a bruise or two?
Not dead? Well then, stuff happens. Don’t complain.
And Tee, enn, ten, vee, five, seven, six, two
Stays on our roads till he kills me. Or you.

(Any resemblance to real-life incidents is not coincidental)

Sequencing the human genome… again

The CSIR is back in the news, this time for more pleasant reasons: sequencing the human genome. Several news items appeared on the achievement, by IGIB, Delhi, last week.

But not all the coverage is positive: as Arvind pointed out in a comment on an earlier post, some senior scientists (including Pushpa Bhargava) question the importance of the achievement as well as the ethics of announcing it to the media before it has been published in a peer-reviewed journal.

I agree on the latter point, and am unsure of the former — but that is the job of peer-reviewers. It is how science works. The media is not qualified to evaluate new scientific claims. The peer-review system, as it currently exists, has its problems, but replacing qualified reviewers by journalists is hardly the solution (I assume that the work has indeed been or is being submitted to a relevant journal, though).

But there were a few other aspects of the original news report that left me disturbed. The comparison of time frames — six weeks for this project, 13 years for the original Human Genome Project — is quite inappropriate. It is always easier to do something for the second time. The original project was developing new sequencing technologies and computational algorithms that took the major part of those 13 years; eventually, the successful method (and one that is widely used today, including, I expect, by the IGIB group) is called the “shotgun” method, where many overlapping fragments are sequenced and then assembled like a jigsaw. These fragments are about 30 nucleotides long, while the human genome has about 4 billion nucleotides in it. Moreover, the genome is highly repetitive, and many of the short fragments would be identical and it would be hard to correctly “assemble” them. Worse, sequencing the short fragments is itself not an error-free method: one or two errors per fragment are expected. So when Celera Genomics came up with the method, they encountered considerable skepticism. Nevertheless, it proved to be the most feasible approach. To alleviate the problem of repetitive regions and sequencing errors, every part of the genome is “covered” 20-30 times by multiple fragments. Even so, completing the assembly for a new organism is a tedious and error-prone process requiring sophisticated software and much human judgement. The point is that these problems are now much better understood than when the Human Genome Project undertook its task, and software is continually getting better. If the IGIB team made significant algorithmic or technical innovations, hopefully they will be described in an upcoming paper.

Even more importantly, the task of sequencing a new human is much easier than that of sequencing a previously unsequenced organism, because a reference genome already exists, and the variation between different humans would be expected to be very small. As I wrote in a comment in reply to Arvind above: one can compare it with assembling a jigsaw with a few billions of pieces, many of which are identical or almost identical, without knowing the “big picture”; versus assembling it with the big picture available to you, knowing that there are only minor differences from the “reference picture”. Technologically, there is nothing very hard any more about this. Equipment and software is marketed for the purpose by large biotech companies like Illumina, and is in use all over the world. If the IGIB team has made significant technical innovations, that is of interest, but it has not been mentioned in news items and it should, of course, be peer reviewed before it hits the media.

As for possible medical benefits: similar claims were made in support of the original Human Genome Project, but little benefit is seen so far. But these things should be seen as basic research, with medical benefits a possible and welcome spin-off, but not the primary goal. It is not at all easy to “link” specific genetic variations with specific diseases, and sequencing a handful of new genomes will not, I think, directly aid that problem. So what is the primary scientific accomplishment here? The news items don’t say, but then, they should not be the primary medium of communicating this work. I look forward to the peer-reviewed article, when it appears.

Jazzer, drop your sax, it’s Jazz Police!

Did Leonard Cohen foresee this?

Why can’t we all just listen to Jug Suraiya?

Jug Suraiya, the Times of India’s self-appointed humorist and contrarian, takes a hatchet to climate science today. He complains that climate change is treated as an irrefutable “fact”. He claims that anyone who tries to question climate change is “immediately branded a destroyer of the planet, a dangerous heretic who should be burnt at the stake” and the “warmists” “will not tolerate sceptics any more than did the Spanish Inquisition”.

So what evidence does Mr Suraiya offer against the hypothesis of manmade climate change? None whatever! He resorts to citing “dissenters”, but names only two, whom he calls “the most notable” — but neither of them is a climate scientist. The first is a geologist, Ian Plimer: read about him and his error-riddled book here and here. And his response to being challenged on his facts by George Monbiot is here. And Mr Suraiya’s second authority is not a scientist at all, but a journalist, Christopher Booker.

Future climate-change deniers will now be able to cite a third authority, the eminent Indian humorist and contrarian, Mr Jug Suraiya. It is a form of proof by mutual reference.

So here are two things to understand, Mr Suraiya. First: man-made climate change is not an “irrefutable fact”. It is a hypothesis for which there is, at this point, an enormous amount of evidence compiled by climate scientists around the world. To refute the hypothesis, you need a significant amount of contrary evidence. Plimer’s book does not cut it.

Second: you allege that big money — “huge money”, in your words — is promoting the climate change hypothesis. In fact most climate scientists work at respected universities and government organisations, on public funding. If you want to follow the money, take a look at who is funding the deniers.

Shiva Ayyadurai rumbles on…

UPDATE 7/12:I notice the folks at the Shiva Ayyadurai Fan Club have linked to my post as alleging a nexus between Nature India and Shiva Ayyadurai. I want to clarify that I am alleging no such thing. I am only saying that they seem to have swallowed one side of the story without asking questions, and given Shiva Ayyadurai (a questionable character) space to vent his spleen on their website without allowing adequate response from CSIR; and they now have some uncomfortable questions to answer.




In my previous post on this subject, I referred to unsubstantiated allegations about Shiva Ayyadurai’s unethical behaviour that I had heard. Subsequently, two CSIR scientists, Vinod Scaria and Sridhar Sivasubbu, wrote the same accusations in two Nature Forum posts. Astonishingly, those posts have been removed by the forum administrators: it seems Nature is intent on promoting Shiva Ayyadurai’s version of the story and will brook no dissent. However, Scaria’s and Sivasubbu’s versions are archived here and here respectively — as of this writing, they look the same to me as the ones that used to be on the Nature forum.

Also, from this post (see comment), it seems that the Nature editors (in London, not India) objected to personal accusations made in those and other posts. I wonder, then, why they agreed not only to publish, but to highlight, accusations made by Shiva Ayyadurai in his article (which continues, as I write, to be frontpaged in the Nature India website) — accusations which include fraud, financial wrongdoing, and arson to cover up the wrongdoing.

For more entertainment, read the rest of the Shiva Ayyadurai blog.

I think this episode is a disgrace and a blot on Nature’s record. (As also the New York Times and others who have given this fraud and sleazeball a pulpit.)

And Nanopolitan has 182 comments and counting. I haven’t yet waded through all this.

But, once again, I would like to ask Prof Samir Brahmachari: why was this creature appointed to CSIR-TECH in the first place, and in what capacity was he appointed?

And while CSIR does do some outstanding science and includes some world-class laboratories, there is no doubt that it would benefit greatly from some changes in structure and management, and I hope some well-intentioned, honest, capable and qualified people are already working on it, without seeking their 15 minutes of fame. There are lots of such people in CSIR already (and elsewhere in India). In fact, I think the state of Indian science (including CSIR) is getting better, not worse, and while there is need for further improvement and change, there is no need for panicked reactions.