String research down 33% in one year-ideas why?

[marcus]

String research down 33% in one year--ideas why?

The index here uses the Harvard abstract database with these six
keywords = superstring, M-theory, brane, compactification, heterotic, AdS/CFT

It counts peer-review publication of professional journal articles. How it changes from year to year helps gauge research activity.

Looking just at the first three months of each year we get the following publication counts:

2002 1340
2006 1442
2007 1433
2008 952

For the first four months of each year,
2002 1596
2006 1800,
2007 1781
2008 1179

For the first six months
2002 2212
2006 2537
2007 2513
2008 ... ?

The data aren't in for the first six months of 2008 so here is the link.
http://adsabs.harvard.edu/cgi-bin/n...txt_wgt=YES&ttl_sco=YES&txt_sco=YES&version=1
When we check this link in July it will give the count.

This is something of a mystery. What could be the cause of this puzzling abrupt change?
I have watched this kind of publication index over the years and it has always been slowly changing. If you use the same keywords year after year the annual change is small. Not anything like 30%. So this should make you curious.

Has there been a computer failure at the Harvard library? A general strike of librarians? Have these six keywords abruptly become unfashionable? Does this correlate with other indices of research health or is it just an isolated oddity?

The way we test our understanding is by predicting. In this case what do you think the index will be for six months? After that, we can try for the whole year. My guess is that there are some fundamental reasons which will have a long-lasting effect. Perhaps, if anyone reads this and takes an interest, they can suggest some.

BTW I hope nobody is naive enough to jump to the conclusion that research stats are a way of determining the ultimate validity or invalidity of extradimensional stringy investigations.
That's something Nature will presumably decide after there is an explicit theory that makes definitive new predictions.

So, for starters, what's your guess about the first six months of 2008?

 

[marcus]

a partial explanation has already appeared!
The Harvard librarians were slow entering all the titles for the first four months into the database. So when I checked today, instead of 1179 we now have a count of 1220, for the first four months of 2008.
So the decline is more moderate
For the first four months of each year,
2002 1596
2006 1800,
2007 1781
2008 1220

 

[Chronos]

The landscape debacle has dampened interest in string theory. I find this interesting. The fundamentals of ST have been used to demonstrate it is unprovable.

 

[marcus]

Betwixt,
The issues around the String Landscape are somewhat sensitive---could be a bit of a sore point. It wasn't my intention to discuss it here.

At the Harvard data base there are always STRAGGLERS.
Publications which are entered late, after the time period is over. So that can bring the counts up. I just checked and the decline (for the first four months of the year) now shows as 31 percent from 2007 to 2008, rather than the 33 percent I said earlier.

We aren't trying to be terribly precise in any case. The puzzle is how did this substantial decline come about?

If you look at the poll, you can see two ideas of how it happened represented. One idea is that it is just a random fluctuation so that the rest of the year will bring the average back up and the final total will be about the same as last year.

How you understand it is show by what you predict.

The other idea is that there is something going on that is behind it, and that it will affect the whole year.

A possible thing could be that some of the more productive people, who write a lot of papers, have gone over into new fields and aren't doing string any more. We know of individual cases which are anecdotal evidence of this. That still leaves a whole lot of people in the field, but maybe they are the ones with fewer ideas and who write fewer papers. So there is less steam. But that is only one of several possible mechanisms you could suggest, I just offer it as an example of a mechanism that could cause a secular trend. Maybe someone else will suggest some other possibilities.

 

[DavidWhitbeck]

With LHC finally approaching completion, funding in high energy theory is probably switching over from string theory to phenomenology. Less funding means less people working on string theory, means less publications.

Also Smolin's book and Woit's book each had a negative impact on the publics view of string theory, and that could also have an impact on funding.

I bet it's a funding issue. But this is just speculation, I don't have any cold, hard facts.

 

[marcus]

With LHC finally approaching completion, funding in high energy theory is probably switching over from string theory to phenomenology. Less funding means less people working on string theory, means less publications.
...

David, thanks for tossing out those suggestions. I think the LHC could very well have something to do with it, and shifts in funding as well (whatever the connection, if any, between those two factors.)

In fact I saw a report last year by the HEPAP---the panel that advises government agencies on high energy physics funding---which said that a survey of university departments showed that they planned to cut back string faculty positions by 15 percent and increase phenomenology positions by around 10 - 15 percent between now and 2012. The biggest planned increase was in astrophysics however (not phenomenology), as I recall. Shifts both in funding and hiring could be having an effect.

If those sorts of reasons are responsible then I suppose one should look at this as probably not just a random temporary dip but as a trend likely to persist somewhat over time.

In any case we can make some crude distinctions and check hypotheses by predicting. You are cordially invited to take a guess, David. We have four predictions so far.

 

[marcus]

Again, my apologies, but by String Landscape are they referring to the... "fabric" that the many strings create??

No apologies needed! the "Landscape" refers to the vast plethora of different versions of physics that one can get out of stringy models---with no selection mechanism in sight.

It used to be thought that string thinking would lead to a unique result, depending on at most one chosen parameter, the string tension, and that would explain why the masses of the particles and the strength of the forces were exactly what they are. Some papers in the 1990s raised doubts about this, but optimism prevailed. Then in 2003 a paper by four authors with initials KKLT appeared which showed that there are at least 10⁵⁰⁰ different string versions of physics, or "vacua". some such unimaginably large number. Predictivity went out the window.

Essentially KKLT found that string could give you anyone of 10⁵⁰⁰ universes because there were so many different ways of rolling up the extra dimensions. the extra dimensions proved a liability. this is an over simplification but not completely wrong either.
and people argue about the exact size of this very large number I called 10⁵⁰⁰ .
(Think of 1 followed by 500 zeros.)

So the answer to your question about "fabric" is no. The landscape is not a fabric. It is a haystack of imagined possibilities in which the needle of reality may or may not be hidden---no way of telling if it is or not.

 

[DavidWhitbeck]

The biggest planned increase was in astrophysics however (not phenomenology), as I recall.

As in particle astrophysics such as Ice Cube or Amanda?

 

[marcus]

As in particle astrophysics such as Ice Cube or Amanda?

The HEPAP report didn't do a breakdown. You are right that there would be a lot of particle astrophysics. Cosmology would also figure in. There is considerable research interest in dark matter and dark energy---do they exist or are there alternative explanations, if they exist what are they? Microwave background mapping and the early universe. Gammaray astronomy is growing---orbital instruments and also groundbased Cherenkov imaging. So-called UHECR astronomy (ultra-high-energy-cosmic-ray). Gravity wave astronomy. Large arrays of various kinds. A lot of basic physics questions---what causes these beyond-supernova GRB (gammaray burst) explosions? what accelerates cosmic rays and how do the interact with the CMB and intervening medium? The physics of stellar collapse and very compact stars. As you indicate, neutrino astronomy too!

There is a lot to try to summarize. Basically astrophysics is a branch of physics more or less coextensive with astronomy and cosmology which is rich both in EXTREME PHENOMENA and NEW INSTRUMENTS (both orbital and ground based). So there is a lot of research opportunity.

If you are interested in seeing what HEPAP had to say you could probably find the report with Google. It came out summer 2007 as i recall. If you try and have trouble finding it let me know.

 

[marcus]

... With so many possibilities, and the for the reason those possibilities can't be avoided, I can see why it would now become a trend for a decline in submissions to occur.

Wouldn't the selection mechanism be the the theory that brings us closest to the Big Bang, though?

Researchers have various personal ways of deciding what to research, and the funding agencies allow for some latitude. So there is no one criterion or guide or focus.

there is a lot of inertia too. People tend to keep on doing the kind of research that they already have experience with and know how to do, or they were trained to do.
It may get harder to find anything interesting in their specialty---they may slow down, or produce fewer and lower quality papers. But they don't always jump out of a declining field.
================

What you say about the big bang raises an interesting question. The big bang is shorthand for the standard cosmology model---a model with expanding distances, that expand according to a simplified version of Einstein's equation called the Friedmann equation.
There is a broad consensus about this model cosmology. It fits the observational data remarkably well. It keeps being consistent with new data as it comes in.

But the initial singularity is a FLAW in that model. It works very well up to but not including the start of expansion and then it doesn't compute. It breaks down.

So the question is how to IMPROVE the standard cosmology model to get rid of this breakdown and be able to run it back prior to the start of expansion.
==================

So instead of preferring theories which "get close" to the Big Bang what I would say is I would like to see a theory which basically does two things:
A. gets rid of the singularity and runs on back to before expansion started----like maybe there was a contraction and a bounce, or something else.

B. explains the interaction of matter with geometry, how matter is connected with geometry so that it can bend it----how matter and geometry arise from the same primitive microscopic description.

You could say that what I want is the same as what you do---that my A. and B. would be a way to "get close" to the Big Bang, because they would in effect let us understand it.

 

[marcus]

Hey we have four predictions entered in the poll already!

Thanks to Arivero, NerfMonkey, and Betwixtwists for registering what they expect. We have quite a spread of forecasts.

this is actually a remarkable moment in contemporary research history----I've never seen such a sharp drop-off as we had in the first four months, so there is a real question about what follows. Was it just a random blip that evens out, or is it a lasting trend.

I'm in some suspense about this.
In case you want to get some kind of early indication, here are the FIVE-MONTH benchmarks:

2002 1939
2006 2157
2007 2187
2008 ...

these are the peerreview publication figures for the first five months of each year with the same six keywords we've been using.

To the extent that you think the whole year 2008 publication should be roughly on par with last year, then you might conjecture that the same would be true of the first five months. So that would be something to check. And would help predict.

 

[marcus]

I think we have allowed enough time for stragglers to come in so we have a roughly stable picture for the first four months of 2008---I'll post the figures.

What we see after late registrations are tallied is still a shocking drop-off. Something like 27 percent in one year. If the rest of the year bears this out it will certainly draw comment and be asking for explanation.

This is a year when (as one of the other posters observed) one would expect theorists to be "working their tails off" deriving and publishing predictions from their various theoretical models so that they might be shot down or confirmed by the LHC when it gets properly going in 2009.

Anything published after 2008 will be apt to seem like a "post-diction". Rarely as convincing.

So this is a surprising anomaly. It should be a banner year for fundamental particle theory publication, and it is turning out the opposite (at least in the stringy category we're looking at.)

Anyway here are the links and data
For the first four months
2002: http://adsabs.harvard.edu/cgi-bin/n...txt_wgt=YES&ttl_sco=YES&txt_sco=YES&version=1

2006: http://adsabs.harvard.edu/cgi-bin/n...txt_wgt=YES&ttl_sco=YES&txt_sco=YES&version=1

2007: http://adsabs.harvard.edu/cgi-bin/n...txt_wgt=YES&ttl_sco=YES&txt_sco=YES&version=1

2008: http://adsabs.harvard.edu/cgi-bin/n...txt_wgt=YES&ttl_sco=YES&txt_sco=YES&version=1

1603, 1806, 1799, 1319

You can see the 27 percent drop from 1799 (the 2007 figure) to 1319 (this year.)
Only partial data is in so far for the first five months but you can see how it is going by clicking the 2008 link below here.

2002: http://adsabs.harvard.edu/cgi-bin/n...txt_wgt=YES&ttl_sco=YES&txt_sco=YES&version=1

2006: http://adsabs.harvard.edu/cgi-bin/n...txt_wgt=YES&ttl_sco=YES&txt_sco=YES&version=1

2007: http://adsabs.harvard.edu/cgi-bin/n...txt_wgt=YES&ttl_sco=YES&txt_sco=YES&version=1

2008: http://adsabs.harvard.edu/cgi-bin/n...txt_wgt=YES&ttl_sco=YES&txt_sco=YES&version=1

1946, 2163, 2199, ...

 

[marcus]

the bulk of the data is now in for the first five months and we can see how it is going by clicking the 2008 link below here.

2002: http://adsabs.harvard.edu/cgi-bin/n...txt_wgt=YES&ttl_sco=YES&txt_sco=YES&version=1

2006: http://adsabs.harvard.edu/cgi-bin/n...txt_wgt=YES&ttl_sco=YES&txt_sco=YES&version=1

2007: http://adsabs.harvard.edu/cgi-bin/n...txt_wgt=YES&ttl_sco=YES&txt_sco=YES&version=1

2008: http://adsabs.harvard.edu/cgi-bin/n...txt_wgt=YES&ttl_sco=YES&txt_sco=YES&version=1

1946, 2163, 2202, 1461 (incomplete)

The number 1461 can be expected to increase some over the next few weeks, as more titles are registered. Even allowing for that, it's a remarkable drop, on the order of 30 percent just from last year.

I would be glad to hear whatever explanations anyone can think of for this.

Other kinds of shrinkage are occurring as well. The annual conference Strings 2008 (at Cern) will be considerably smaller, judging by the posted program, than earlier conferences such as in 2005 and 2006. Fewer string theory talks this time.

The Advisory Panel on High Energy Physics to the main funding agencies NSF and DOE reported last year that university departments plan a 15 percent cutback in the number of string faculty jobs in the US between now and 2012---another kind of shrinkage. (2007 HEPAP report)

So what is going on? And are these things related? Any ideas?

 

[marcus]

Here's an update on how stringy research publication is going this year.
I am using six keywords at Harvard Abstracts:
superstring, M-theory, brane, AdS/CFT, compactification, heterotic.

These are the publication numbers in successive years
2002, 2006, 2007, 2008

For the first four months of each year
1603, 1806, 1803, 1325

And for the first five months of each year
1946, 2163, 2203, 1493 (may increase with late entries)

Here are links where you can check yourself if you want. The numbers can change as the librarians find more stuff to add to their database. But they tend to stabilize.

2002: http://adsabs.harvard.edu/cgi-bin/n...txt_wgt=YES&ttl_sco=YES&txt_sco=YES&version=1

2006: http://adsabs.harvard.edu/cgi-bin/n...txt_wgt=YES&ttl_sco=YES&txt_sco=YES&version=1

2007: http://adsabs.harvard.edu/cgi-bin/n...txt_wgt=YES&ttl_sco=YES&txt_sco=YES&version=1

2008: http://adsabs.harvard.edu/cgi-bin/n...txt_wgt=YES&ttl_sco=YES&txt_sco=YES&version=1

Any ideas? Explanations?
======================

There are two things you might want to consider as helping explain.

A. there is a rapidly growing worldwide research movement in background independent quantum gravity
with a wealth of new results (since 2004, and even in just past 3 months), which just began holding annual interanational conferences in 2005, 2007, 2008
it combines approaches called LQG, CDT, spinfoam, LQC, noncommutative geometry (NCG) also called quantum geometry.
this might be relevant to what we are seeing in stringy research, or it might not be relevant.

B. stringy research seems to be stuck, after an exuberant enthusiastic period up thru about 2002 there haven't been much new results or highly cited papers
string research seems to be bogged down in a "multiverse" of possibilities, with no way to make unique predictions about new observable physics. there have been expressions of discouragement from leaders such as David Gross. Edward Witten has redirected his research mostly out of string and into other lines of investigation. he has expressed dissatisfaction with the multiverse of stringy versions of physics.
It could be that many of the best people have moved over into other fields. that could also possibly be relevant to the decline in publication numbers. (although it creates room in the field for less talented individuals, who can fill in the gaps)

 

[marcus]

We now have 5 predictions on record.

I've already clearly lost!

It looks like a race between Nerfmonkey and Betwixtwists.

We are coming down on the 6 month mark, so it should be clear within a few weeks which forecast is closest to being right. Maybe 3 weeks to allow for stragglers.