In an earlier post I went over changes in word usage in philosophy articles. Here I want to do something a bit simpler. I’m going to graph out how much longer articles are getting.
The wordcounts that I get from the JSTOR databases and pdftools are a bit longer than what you’re probably used to, because they include everything. That means the body text, the footnotes, the bibliography, the abstract, the running head, and so on. In practice, that means they are about 10-20% longer than what you’d normally think if you were just looking at body text and footnotes.1
1 So for some of you, that means about 100% longer than body text alone.
The things like the running head are annoying to include, but they haven’t changed much over time. So the trends in the graphs here would be the same even if we used a more normal measure.
We’ll start, in Figure 1, with a look at all twenty journals.2 I haven’t used mean lengths, because it can get thrown off too much by single long articles. Instead, for each year I’m plotting the 25th, 50th, and 75th percentile word lengths. So the lowest line is the length such that a quarter of articles are shorter, the middle line is the length where just as many articles are longer than it as shorter, and the top line is the length such that 25% of articles are longer.
2 See the earlier post for the list of journals.
3 The lines here are five year moving averages, with the window centered on the year in question. So above 2006, the lines show the 25th, 50th, and 75th percentiles for the window 2004-2008.
I won’t go over it here because it’s a bit of a mess, but I’m mostly taking Web of Science’s classification of what’s an article at face value. This excludes book reviews, and it excludes most things listed as ‘Discussion’ or ‘Note’ in the journal, and it’s somewhat random whether it includes or excludes book symposia and critical notices. (They don’t seem to do this consistently, and I don’t know a good way to correct it.) Including any of these would bring the lengths down a bit. That said, Figure 1 is the big story.3
The trends here are remarkable. In every year from 2015-2019, the 25th percentile article, by length, was longer than the median article in any year from 1980-1984. There is a little less movement at the top end, but the disappearance of short articles is rather stunning.
You can see more or less the same trends across all of the journals in Figure 2. The dots here bounce around a lot more, and even the moving averages are fairly noisy.
There are some journals which look like the overall trend. In this respect, Philosophical Studies looks most like the discipline as a whole. (This is true in several other respects as well, but that’s a story for another day.) But the steady march upwards we saw in Figure 1 isn’t always apparent in Figure 2. What is somewhat more common among the individual journals is compression. For several journals, the gap between the three lines is much closer in 2019 than it was throughout the previous four decades. There are some journals where that gap has stayed steady; the only one where it has really expanded is Philosophy of Science.
There is a natural story to explain this. Journals which previous just published articles at the length that people wrote them had to start imposing word limits, and then enforcing them tightly. As philosophers wanted to write longer and longer articles, they started sending papers to journals where they were just under the limit. And so we get situations like with the Australasian Journal of Philosophy where the word limit is tightly enforced, but practically every published article is butting against that limit.