Archive for Avi Loeb

Inflationary Perturbation

Posted in The Universe and Stuff with tags , , , , , , on May 11, 2017 by telescoper

I thought I’d just draw the collective attention of my vast readership (Sid and Doris Bonkers) to a bit of a row that has broken out between two groups of cosmologists concerning the theory of cosmic inflation.

This kerfuffle started with an article entitled Pop Goes The Universe in Scientific American by Anna Ijjas, Paul Steinhardt, and Avi Loeb that (among other things) asserts that inflation “cannot be evaluated using the scientific method” and is consequently not a scientific theory. Another group of cosmologists (including Alan Guth, the author of the paper that launched the inflationay universe model) penned a response that was signed by a long list of leading scientists, thirty-three of them to be precise. The original authors then issued a response to the response. Sean Carroll (who was one of those who signed the response the original paper has written a nice blog post summarizing the points of disagreement.

I’m not going to attempt to post a detailed response to every issue raised in this correspondence, but I will make a few points.

First, I think it’s important to realize that there isn’t a single simple definition of `the scientific method’: there are lots of scientific methods, each of which may employed to a greater or lesser degree in different disciplines. Most scientists would probably agree that some notion of `testability’ has to be included if a theory is said to be scientific, but it seems to me that testability is not an absolute, in the sense that not all predictions of a theory need to be observable for the theory as a whole to be testable to a degree. A theory might predict the existence of a phenomenon A that is impossible for all practical purposes to observe, but if that theory also has another necessary consequence B that is observed then the theory does not deserve to be dismissed as unscientific.

One aspect of modern inflationary theory that is singled out for criticism has been the incorporation of the idea of a multiverse. I have to make the confession here that I don’t like the concept of the multiverse, nor do I like the way it has become fashionably mainstream in the field. I’ve never seen it as a necessary (or even useful) addition to inflation theory. However, suppose you have a model of inflation that leads to something like Linde’s version of the multiverse. Causally disconnected domains of this multiverse may indeed not be observable, but if the theory has other necessary implications for things we can observe in our local universe then it is testable to a degree.

My position (such as it is) is that I like the idea of inflation, largely because: (a) it’s very neat; and (b) it provides a simple mechanism for generating fluctuations of the right form to account for so many of the observable properties of our universe, especially the fluctuations we measure in the cosmic microwave background seen by Planck:

These observations don’t prove that inflation is right, nor do they narrow down the field of possible inflationary models very much, but they do seem to be in accord with the predictions of the simplest versions of the theory. Whether that remains true for planned and future observations remains to be seen. Should someone come up with a different theory that matches existing data and can account for something in future data that inflation can’t then I’m sure cosmologists would shift allegiance. The thing is we don’t have such an alternative at the moment. Inflation is the preferred theory, partly for want of compelling alternatives and partly because we need more data to test its predictions.

That said, there are one or two points on which I agree with Ijjas, Steinhardt and Loeb. In particular there has developed what I consider to be a pathological industry dreaming up countless variations of the basic inflation model. There is now a bewildering variety of such models, few of which have any physical motivation whatsoever. I think this is a particularly a grotesque manifestation of the absurd way we measure scientific `success’ in terms of counting publications and how that has driven unhealthy research practice.

No doubt many of you disagree or wish to comment for other reasons either on the original communications or on my comments. Please feel free to offer your thoughts through the box below!

Building Blocks and Blueprints in Cosmology

Posted in The Universe and Stuff with tags , , , on August 1, 2013 by telescoper

Still playing catch-up from my recent travels, so to provide a blog post for today I’ve decided shamelessly to rip off an interesting comment on a blog post by Sean Caroll which picked up on the theme I posted about a few days ago, namely my perception that the current generation of cosmologists seems rather reluctant to question the standard paradigm. Please bear with me if that all sounds a bit incestuous…

Anyway, Peter Edmonds commented in order to draw attention to a series of papers on related matters by Avi Loeb (of Harvard University) which can be found on the arXiv here, here, here and there.
I’d encourage you to read the four interesting papers I’ve linked to above as I think they are extremely thought-provoking. The last of these begins with this paragraph, so you can see why it’s relevant to the aforementioned topic.

Too few theoretical astrophysicists are engaged in tasks that go beyond the refinement of details in a commonly accepted paradigm. It is far more straightforward today to work on these details than to review whether the paradigm itself is valid. While there is much work to be done in the analysis and interpretation of experimental data, the unfortunate by-product of the current state of affairs is that popular, mainstream paradigms within which data is interpreted are rarely challenged. Most cosmologists, for example, lay one brick of phenomenology at a time in support of the standard (inflation+Λ+Cold-Dark-Matter) cosmological model, resembling engineers that follow the blueprint of a global construction project, without pausing to question whether the architecture of the project makes sense when discrepancies between expectations and data are revealed.

To put this another way, a great deal of modern astrophysics and cosmology is rather incremental. I don’t mean that in a derogatory way, just that such research often involves large-scale observational projects that have to proceed slowly and painstakingly. Working at the coal face in large consortia like this makes it difficult to take the time to step back and consider the bigger picture. We ask a lot of early career researchers nowadays when we expect them to cope with detailed analytic work as well as assimilating and synthesizing a coherent view of the overall landscape. Producing a stream of research papers doesn’t in itself make an excellent research. Productivity needs to be balanced by a proper appreciation of which questions are the most important ones to ask, which often requires (and I apologize for using such an awful cliché) thinking outside the box.