There’s a report in Science dated June 8 with the headline ‘Middleweight black holes found at last’. The abstract describes an effort by an “international team” of astronomers to find intermediate-class black holes, which weigh more than tens of solar masses but less than millions of solar masses.

This is a bit confusing because labels in the natural and social sciences have fixed and well-defined meanings. So “intermediate-class” means something specific. When the first LIGO announcement of a black hole merger was made in February 2016, Karan Jani, a member of the LIGO Scientific Collaboration, told me that “intermediate-class” black holes weighing 20-10,000 solar masses generate the loudest signals in the instruments.

Doesn’t this mean black holes fitting the label intermediate-class were found in 2016 itself?

Further, this isn’t even the second time intermediate-class black holes have been announced discovered. In February 2017, astronomers from the Harvard-Smithsonian Centre for Astrophysics announced that they’d found a black hole weighing ~2,300 solar masses at the centre of a globular cluster called 47 Tucanae.

Another thing that concerns me here is how the data is being sliced. You have astronomers claiming discoveries every day and, given a recent spate of articles about the neutron-star merger discovery and the BICEP2 ‘cosmic blunder’, you know astronomy and cosmology are ultra-competitive realms of scientific endeavour. As a result, in such cases, there is often a real risk that someone out there will claim to discover something that is not really significant at all.

For example, it would seem intermediate-class black holes have been discovered thrice (LIGO, 47 Tucanae and when the “international team” mentioned above discovered them at the centres of 300 galaxies). At this rate, it’s quite possible this kind of black holes has been discovered even more. Which one was the first one? Or is the ‘intermediate-class’ itself going to be cut up into three or more parts to accommodate all these claims? Additionally, if you’re wary about using the term ‘intermediate-class’, then you should know that the term ‘middleweight’ also has precedence.

A press release from the Harvard-Smithsonian Centre for Astrophysicists in February 2017 had this line:

Astronomers expect that intermediate-mass black holes weighing 100 – 10,000 Suns also exist, but so far no conclusive proof of such middleweights has been found.

Lee Billings, a science writer for the Scientific American, wrote in June 2017:

Most of the black holes in LIGO’s mergers have been middleweights, being heavier than that 20–solar mass limit but much lighter than the supermassive variety, raising questions about their origins and relationship to the two well-studied populations of black holes. (emphasis added)

It’s likely that in both these cases the authors are using the term ‘middleweight’ to refer to the intermediate class in a non-technical way – but then so is the Science article.

Everybody remembers the infamous case of Brian Wansink, who tortured his data and sliced his results into smaller and smaller pieces that he published as separate papers. Nobody wants that sort of thing to happen because what Wansink did wasn’t science; he was simply hacking the system for personal gain. However, if we can’t reach consensus on what intermediate-class really means and, following that, when the first black hole of this class was found, we might never see an end to scientific papers and research groups claiming that their authors/members have discovered a new class of something.

There are a few issues that could keep such consensus from being reached; I can think of three.

1. The intermediate-class comprises four orders of magnitude: hundreds, thousands, tens of thousands and hundreds of thousands of solar masses. I’m no astronomer but this still sounds like a very wide swath to homogenise, especially if professional astronomers are going to claim the discovery of a black hole of each order of magnitude is significant. (E.g. the abstract of the latest discovery assumes intermediate-class black holes weighs thousands to millions of solar masses, excluding the hundreds.)

2. LIGO used gravitational waves to spot middleweight black holes; the 47 Tucanae group used radio and X-ray data from large observational studies; the “international team” used galaxy spectra data collected in the archives. Since each group has used a different technique to find what they have, does each effort get to stake its own, distinct claim to primacy?

3. The Wansink outcome – where astronomers are slicing the data really thin with the (reasonable) assumption that differences between one supernova and the next are the result of distinct processes instead of a common process with stochastic fluctuations. Of course, we’ve no way until much later in the future, after we’ve made lots more observations, to tell whether the way we’re slicing the data in 2018 actually makes sense. But in the same vein, there should be a way to tell if, based on observations made in the last few decades, we’re classifying black holes right.

(To be clear, while the first and second issues described could imply that astronomers are doing something wrong, the third doesn’t; it’s just something to consider.)