Further Reflections on the Most-Cited Works in the Stanford Encyclopedia of Philosophy: Underranked Works and Concentration Percentage

A couple of weeks ago, I published a list of the 253 most-cited works since 1900 in the Stanford Encyclopedia of Philosophy. (The SEP had 1778 main-page entries as of my scrape last summer, and many of those entries have long reference lists.) Citation in the SEP is plausibly a better measure of impact in mainstream Anglophone philosophy than other bibliometric measures like Google Scholar and SCOPUS, which include citations by non-philosophical sources (which can dominate citations within philosophy, since philosophy is overall a relatively low-citation field) and which mix citation by sociologically elite venues with citation by less elite venues (and those citation patterns can be very different).

I think informed readers will tend to agree that the works near the top of the list (Rawls' Theory of Justice, Kripke's Naming and Necessity, etc.) are indeed among the most influential works in the mainstream Anglophone tradition -- more influential in mainstream Anglophone philosophy than, say, Foucault's Discipline and Punish or Popper's Logic of Scientific Discovery, despite Foucault's and Popper's higher citation overall across all disciplines and sources.

(What do I mean by "mainstream Anglophone philosophy"? I mean philosophy as practiced by professors in departments highly ranked in the Philosophical Gourmet Report, as published in journals that are highly ranked in Brian Leiter's polls (e.g., here), and -- though this would be circular for present purposes -- as recognized in the Stanford Encyclopedia of Philosophy. Even readers who dislike the philosophy of this tradition, or who see it as troublingly narrow, can I think recognize the sociological phenomenon of influence in these related ecologies, reasonably called "mainstream" in Anglophone academia.)

Underranked Works

Although SEP citation rates are, I think, a better measure of impact in mainstream Anglophone philosophy than any other existing bibliometric measure, that doesn't mean they are perfect. Works with a huge impact on a subdiscipline, or on a particular topic, will plausibly be underranked compared to works with substantial impact across a range of areas. The SEP will have only a limited number of entries for each subdiscipline or topic, and no matter how important the work is to that subdiscipline or topic, it can appear only once in each entry's bibliography.

This explains, I think, the relatively weak showings of some of the best-known articles in the field. For example:

• 119th (tied), 21 citations: Gettier, Edmund L., 1963, Is Justified True Belief Knowledge?

• 192nd (tied), 17 citations: Anscombe, G.E.M., 1958, Modern Moral Philosophy

• unranked, 14 citations: Searle, John R, 1980, Minds, Brains, and Programs

• unranked, 12 citations: Singer, Peter, 1972, Famine, Affluence, and Morality

• unranked, 9 citations: Thomson, Judith Jarvis, 1971, A Defense of Abortion

This isn't intended as any kind of exhaustive or representative list of underranked works -- just a few examples that struck me as conspicuously underranked relative to their influence. Gettier's 1963 article is possibly the most influential work of 20th century epistemology (in mainstream Anglophone circles). Anscombe's 1958 article is often seen as a landmark in the resurgence of virtue ethics. Searle's 1980 "Chinese room" argument is perhaps the most influential work on philosophy of computation and artificial intelligence after Turing. Likewise, Singer's 1972 article on charitable donation (with its famous example of rescuing a drowning child in a nearby pond at the expense of your clothes) and Thomson's defense of abortion (with its violinist example) are known to virtually all mainstream Anglophone philosophers.

Works might also be underranked if the SEP has relatively few entries in their field or subfield. For example, I'd venture that epistemology has relatively few entries relative to its overall influence in mainstream Anglophone philosophy. And although feminism has probably been somewhat more influential in mainstream Anglophone philosophy than philosophy of race, SEP features many more entries on the former than the latter, possibly explaining why some important feminist works appear on the list (e.g., Butler's Gender Trouble at rank #61), while philosophy of race is poorly represented.

Influential authors and ideas might also fail to appear on this list, if the influence is spread among several works. For example, here are the ten most-cited authors who have no individual works represented among the top 253:

John Hawthorne (97 total citations)
Jonathan Bennett (83)
William Alston (77)
Judith Jarvis Thomson (72)
William G. Lycan (71)
Nicholas Rescher (71)
Peter Singer (71)
Ernest Sosa (69)
Jeremy Waldron (68)
Joel Feinberg (67)
Amartya Sen (67)

All of the above are among the top 86 most-cited authors born since 1900. So of course no negative inference about the importance of any individual author is justified by the absence that author's individual works from the works list.

What Percentage of an Author's Citations Are to Their Most-Cited Work?

By comparing my most-cited authors list with my most-cited works list, we can get a rough measure of how much an author's impact is concentrated in a single work vs. spread across multiple works. (Note that the lists are not quite comparable, since the authors list includes only authors born 1900 or later while the works list includes all works published 1900 or later, including works by authors born before 1900.)

Consider, for example, Thomas Kuhn. His Structure of Scientific Revolutions was one of the most influential works of philosophy of the second half of the 20th century. Fittingly, it appears 9th on my list of most influential works. But Kuhn himself appears relatively low on the list of most influential authors: 63rd. Looking at the raw numbers, we can see that 58 entries cite Structure and 71 entries cite any work by Kuhn. Thus, 82% of the Kuhn-citing entries cite Structure.

Contrast this with, say, David Lewis, who is the #1 most-cited contemporary author overall (with 307 entries citing his work) and whose most-cited work, On The Plurality of Worlds, ranks #6 (70 citing entries). For Lewis, 23% (70/307) of the entries that cite him cite his most-cited work.

I can't seem to think of a good name for this number, so I'll have to settle with a bad name: the concentration percentage. Here are the concentration percentages of the ten most-cited contemporary authors in the SEP:

1. Lewis, David K.: 23% (70/307)
2. Quine, Willard van Orman: 32% (69/213)
3. Putnam, Hilary: 24% (45/190)
4. Rawls, John: 76% (127/168)
5. Kripke, Saul A.: 58% (92/159)
6. Williamson, Timothy: 32% (48/152)
7. Davidson, Donald: 21% (31/151)
8. Williams, Bernard: 22% (32/146)
9. Nussbaum, Martha C.: 19% (26/140)
10. Nagel, Thomas: 24% (33/137)

Thus, we can see two clusters: A couple of authors had most of their citation impact through a single work: Rawls (via A Theory of Justice) and Kripke (via Naming and Necessity). The remaining authors had about a third to a fifth of their citation impact through a single work.

Among the top hundred authors, the ten most concentrated are:

Kuhn, Thomas S. (82%: Structure of Scientific Revolutions)
Rawls, John (76%: A Theory of Justice)
Parfit, Derek (71%: Reasons and Persons)
Scanlon, Thomas M. (66%: What We Owe to Each Other)
Kaplan, David (65%: Demonstratives)
Ryle, Gilbert (61%: The Concept of Mind)
Kripke, Saul A. (58%: Naming and Necessity)
Ayer, Alfred J. (54%: Language, Truth, and Logic)
Nozick, Robert (53%: Anarchy, State, and Utopia)
Evans, Gareth (53%: Varieties of Reference)

I confess to being surprised that some of these percentages aren't even higher. For example, I'd have guessed Ryle's impact was more than 61% concentrated on The Concept of Mind.

The ten least concentrated are:

Bennett, Jonathan (16%)
Pettit, Philip (16%)
Harman, Gilbert H. (16%)
Hawthorne, John (15%)
Thomson, Judith Jarvis (15%)
Lowe, E. J. (15%)
Waldron, Jeremy (15%)
Feinberg, Joel (13%)
Yablo, Stephen (13%)
Rescher, Nicholas (6%)

I'll venture a prediction. According to the phenomenon I've labeled "The Winnowing of Greats", the greater your distance from a group that varies in eminence, the greater the difference seems between the most eminent members of that group and the less eminent members. (This is to some extent because you have zero knowledge of most members below a certain level of eminence and to some extent because you overrely on second-hand summaries that highlight a few of the most eminent examples.) If this winnowing phenomenon applies to works as well as to authors, then as time creates distance from our era, all but the most influential works will largely be forgotten -- which will disproportionately favor highly concentrated authors in the historical memory.