1.1 Selecting the Twelve Journals
This is a study about the trajectory of topics across leading philosophy journals. But presumably most people reading this aren’t interested in philosophy journals as such; they are interested in the trajectory of philosophy. So it is important to select, as far as possible, journals that accurately reflect what’s going on in philosophy.
An obvious idea would be to just use generalist journals, because they will reflect what’s generally happening in philosophy. But this turns out to be a bad idea, since there really aren’t any generalist journals in philosophy. Perhaps that’s because the journals in moral and political philosophy, and in philosophy of science, are so good, that so-called generalist journals tend to under-represent work in those fields. Or, perhaps more precisely, they don’t always reflect the cutting-edge work in those fields.
In previous work, I noted how little attention the leading generalist journals had paid to two of the most important late twentieth-century articles, Elizabeth Anderson’s “What is the Point of Equality?”, and Peter Machamer, Lindley Darden and Carl Craver’s “Thinking about Mechanisms”. These papers each have over 2500 Google Scholar citations, but they have barely been mentioned in the leading generalist journals. An accurate picture of recent philosophy has to include the literatures these papers spawned, and those literatures on the whole aren’t found in generalist journals. So to get an accurate picture of philosophy, you need to include at least some specialist journals.
As a reminder, here are the journals that I’ve included:
|Journal||First Year||Number of Articles|
|British Journal for the Philosophy of Science||1950||1803|
|Journal of Philosophy||1921||4389|
|Philosophy and Phenomenological Research||1940||3300|
|Philosophy and Public Affairs||1971||609|
|Philosophy of Science||1934||3551|
|Proceedings of the Aristotelian Society||1888||2322|
|The Philosophical Quarterly||1950||1905|
As you can see, there are two moral and political journals, Ethics and Philosophy and Public Affairs (P&PA), and two philosophy of science journals, Philosophy of Science and the British Journal for the Philosophy of Science. I could possibly have gotten by with just one of each. But I thought Ethics and P&PA brought in different fields of philosophy, and so they were both worth including. That meant it would be good to balance them with two philosophy of science journals. This had the side benefit of my not having to decide which of those two philosophy of science journals was more representative.
But if I had those four “specialist” journals, I needed enough “generalist” journals that what I had felt representative of philosophy as a whole. Partially to get the balance right and partially to make the graphs look nice, it felt like I needed eight more journals. I started with the current “big four” journals.
- The Philosophical Review
- Journal of Philosophy
I added Analysis because I wanted to be very sensitive to trends, and Analysis is often ahead of the trends in the field. That leaves three more spots. Here were the criteria I used to fill those:
- The data for the journal had to be available through JSTOR’s Data for Researchers. This was not negotiable since that was my data source. But it was unfortunate, since it ruled out the Australasian Journal of Philosophy, which would otherwise have been perfect.
- The journal had to be active for a long time. It wouldn’t help balance much to add a journal that didn’t exist during the timeframe I’m looking at. This ruled out Philosophical Studies. That’s a bit of a shame since the story of twenty-first century philosophy can’t be told without Philosophical Studies. But it just doesn’t have enough history for the purposes of this study.
- The journal could not be too idiosyncratic. I wanted it to tell something about the field, not just about those journals. This ruled out The Monist, which was very idiosyncratic in its early years. In recent years, it is idiosyncratic in a good way; highlighting work the others sometimes overlook. But before World War II it is barely a philosophy journal in any recognizable sense.
- The journal could not be a philosophy of science journal, since the aim is to balance the two philosophy of science journals I have.
- The journal had to primarily publish in English, since the analysis tools I’m using simply don’t work for cross-linguistic data sets. The last two criteria ruled out Synthese and Erkenntnis.
After all that, I was left with:
- Philosophy and Phenomenological Research: This has slightly more non-English work than is ideal for current purposes, but I thought adding a little continental philosophy from its early years was worthwhile. And it became such an important journal that it felt wrong to leave it out.
- Proceedings of the Aristotelian Society: Note that I’m including the supplementary volumes here.4 This is idiosyncratic in its early years; some secretaries of the Aristotelian Society have a bigger impact on this study than they do on the field. But without it, so much of British philosophy is missed, including some themes in contemporary philosophy that aren’t always covered in the other major journals.
- Philosophical Quarterly. For much of the twentieth century, this is much less prestigious than the other eleven journals I’m looking at, and this will become relevant when I look at the citation data. But it fits the other criteria very well. It adds a Scottish journal to the English and US journals I am otherwise looking at. It has slightly better history coverage than the other journals, and since I worked at St Andrews for so many years, I’m personally fond of it.
One might wonder why I didn’t add any other specialist journals, along with journals in ethics and in philosophy of science. The reasons were a bit varied.
I think there is a better sense of midcentury philosophy if one logic journal is added. But text mining can’t be done on symbols. And in more recent years, the sense in which the logic journals are primarily philosophy journals as opposed to mathematics journals has gotten weaker. So, I left them out.
The twelve journals I have don;t include as much history of philosophy as there is in the profession. But that’s simply unavoidable if doing a study based on journals. History of philosophy is primarily a book discipline rather than a journal discipline. This can be seen in the citation data. Pick almost any prominent figure in history of philosophy and odds are that I’ll have several journal articles with more citations than their most cited article. The prominent figure picked will almost surely have several books that are more widely cited than any of my articles. The point isn’t that historians of philosophy are never cited but that they rarely have highly cited articles. Just as importantly, when a history article is widely cited, it usually appears in one of the twelve journals I’ve already included. For this reason, the model that I end up working with does have a lot of history categories. Just remember that the absolute numbers of articles in each of these categories is not representative of how important the categories are in philosophy.
And the other specialist journals are either too new (e.g., Mind and Language, or Linguistics and Philosophy) or representative of too small a section of contemporary philosophy to be worth including. Aesthetics, for example, is an important philosophical field. (And it shows up in the model in an interesting way.) But including the Journal of Aesthetics and Art Criticism in the study would have made it look like aesthetics was 1/13 of the field, and that’s misleading. So I stuck with these twelve.
Including them also causes a few headaches. The articles in the supplementary volumes often have respondents. When they do, the metadata often lists both the original article and the reply article as coauthored. This is bizarre, though even more bizarre is that often the running head on the print article does the same thing. This can totally mess up the calculating of philosophical Erdös numbers which one would probably know if they have calculated them). This study isn’t tracking authors, so it isn’t too painful. But I am using auto-generated citations as a way of picking out articles, and they will sometimes look coauthored when they are really not. I’m not going to try fixing this; it’s just a weirdness that I’ll live with.↩︎