Tuesday, February 02, 2010

Naming and Ranking

Martin Kruskal invented Soliton Waves which were a very important concept in Physics. (NOTE- one of the comments clarifies this statement.)

Rebecca Kruskal (Martin's Granddaughter): Daddy, how come they are not called Kruskal Waves?

Clyde Kruskal (Martin's Son, Rebecca's Father): You can't name things after yourself.

Rebecca Kruskal: Why not?

Rebecca raises a good question. In academia the etiquette has evolved that you simply do not name things after yourself. Why is this? Is it a good thing? How did this tradition get started? Have people tried to name things after themselves? What happens in other endeavors?

On a related topic: if you are asked to give a list of the top items in your field then is it okay to list some of your own?

In THE NEW BOOK OF LISTS by Wallechinsky (spell check wanted me to change the name to Lewinsky) and Wallace they have several lists where an expert in X lists his favorite things in X. Some include their own work:
  1. Johnny Cash's 10 Favorite Country Songs of all Time includes his own I Walk the Line (at number 1) and Folsom Prison Blues (at number 4). To be fair, they are awesome songs!
  2. Oliver Stone's 12 Best Political Films of all Time actually lists 10 movies (films?) but then says Stone Notes: And two more with apologies: 11-12. JFK and Salvador. Because I never thought either could be made, much less be appreciated by a large audience. This strikes me as a good way of doing it- since 10 is the usual number on a list make it 12 and include two of your own and apologize for it. However, the movie JFK was way too long. The point of the movie was made more concisely here
  3. Federico Fellini's 10 All Time Favorite Films include his own 8 1/2 as number 10.
  4. Lucille Ball's 10 Favorite TV Series has as item 10 and of course I Love Lucy.
  5. Charles M. Schultz's 10 Greatest Cartoon Characters includes, at number 1, Charlies Brown and Snoopy.
If I was asked for my favorite theorems and I had one that I thought was reasonable I wouldn't put it on the list. But I would add at the end something like With apologies I include ..... I think it is unfair to compare your own work with others.


  1. There is a trick to get people to name things after you: give them a bad name in your paper or no name. If they have no good way to refer to it, they will just call it Gasarch's Theorem.

  2. (I have never done that before on purpose, I am just mentioning it because I think it would work)

  3. You don't name discoveries after yourself because it's unnecessary (good work will eventually speak for itself) and distasteful.

  4. This comment has been removed by the author.

  5. paraphrasing an argument I think I read in an essay and found convincing:
    When possible, picking SOME descriptive word or phrase has the potential to be a lot more useful to people when they try to understand, remember, and use such results, even if it isn't a perfect description.

    I suppose that naming things eponymously can only be helpful in this respect if you happen to be familiar with that particular person and what kind of work they did. In the long run, maybe the meaning of words tends to last longer than the memory of a specific person.

    "Lovasz Local Lemma" is a compromise?

  6. It medicine it is an axiom that it is a good thing to have a disease named after you (they are called medical eponyms), but that it is a very bad idea to allow a surgical procedure and/or device to be named after you.

    The reason being, that diseases are eternal, but better procedures and device will surely come along, after which everyone will say "I never use the 〈your name here〉 procedure/device anymore, it's terrible".

  7. Uhhh ... just to state the obvious (but is it obvious?) ... the corollary for mathematicians is that it's good to have a theorem named after you (theorems being eternal), but bad to have an algorithm named after you, *unless* that algorithm is provably optimal.

    Is this in fact how it works? Are eponymous algorithms more scarce than eponymous theorems?

  8. In my math for nonmajors class I taught Kruskal's Algorithm.

  9. "Martin Kruskal invented Soliton Waves which were a very important concept in Physics."

    Hard to parse. Kruskal and Zabusky were the first to use "soliton" in a paper (Physical Review Letters 1965) where they discussed soliton solutions to the KdV equation. The idea of a soliton, however, goes back much further to at least the early 19th century. The analytical solution to the KdV equation was provided later by a larger group of researchers.

    So I wouldn't say that Kruskal "invented Soliton Waves." He and Zabusky did coin the word. So I guess for you question it should be why aren't they called Kruska-Zabusky waves?

  10. Anon #8: the Kruskal of Kruskal's algorithm is Joseph Kruskal, Martin's brother. Relevantly to this post, a different minimum spanning tree algorithm, often called the Prim–Dijkstra algorithm, is actually due to Jarník.

  11. Related question: is it ok to teach your own results in class?

    Generally speaking I find this to be in bad taste (unless the result in question is important enough that someone else teaching the same class would almost surely cover it also).

  12. Related question: is it ok to teach your own results in class?

    For one it depends if you are teaching a grad or an undergrad course.

    ...unless the result in question is important enough that someone else teaching the same class would almost surely cover it also.

    That's the criteria we use around here. We teach the material that we believe is important, regardless of provenance. As such, only a few courses contain results from professors in the department.

    To my surprise, students seem to take particular pride when I point out that a specific result is due to a professor at their university.

  13. In the field of Law and Economics, Nobel Economist Ronald Coase made some wonderful and very funny observations. I laugh my head off every time I watch his opening remarks in his Coase Lecture. The lecture (in 6 parts YouTube) also contain a bit about him talking about the "Coase Theorem" which was badly defined by a fellow economist.

    The opening remarks,

    The full speech with time codes,

  14. Kruskal's Algorithm
    Dijkstra's Algorithm
    AVL Trees
    Floyd–Warshall Algorithm
    Knuth-Morris-Pratt Algorithm

    clearly it is good to be an algorithms guy