cormullion’s blog has a deep dive into the history of the asterisk which is lots of fun (and educational too!):

The asterisk has a long history. The first appearance of this simple mark was probably on a cave wall somewhere, but we like to assign inventions to known individuals, so the inventor of the asterisk was: Aristarchus of Samothrace, in about 200 BCE.

I was disappointed that this wasn’t the other Aristarchus, Aristarchus of Samos, the famous mathematician with an interest in astronomy, because ἀστερίσκος means “little star”. The man from Samothrace howrver was a librarian, scholar, critic, and proofreader, who liked to make numerous marks (*) [marginal note: * Like this.] in the margin of texts and manuscripts, like notes, queries, and critical comments.

If you have a long memory or are into typography, you may be thinking “Isn’t there a Keith Hou­s­ton post about this?” There is, but:

Keith Hou­s­ton’s excellent book Shady Characters covers the history of most of the punctuation marks in great detail. But his chapter on the asterisk concentrates entirely on the asterisk’s use as a footnote indicator, and ends more or less here.

There’s a great deal about multiplication at the link, as well as glorious illustrations.


  1. An author owned an asterisk
    And kept it in his den
    Where he wrote tales which had large sales
    Of erring maids and men,
    And always, when he reached the point
    Where carping censors lurk,
    He called upon the asterisk
    To do his dirty work!

    (Ascribed to Stoddard King.)

  2. An excellent find! Thanks for posting it (and for the name-check).

  3. My pleasure, and how could I avoid name-checking Mr. Punctuation?

  4. Since the article didn’t dive into that, what is the definitive history of using asterisks as wildcard characters in computing? John Cowan?

  5. Two points:

    1) The funniest use of unorthodox glyphs to mark footnotes (and iterated footnotes) was in “The Annotated Calvin & Hobbes,” which appeared in the New York Times on January 9, 1994. I still have the original page torn out, and if you have an online subscription (I refuse to give the FTFNYT any money, or even my information), you should probably be able to view a PDF of the page if you search for the the name.

    2) The post on cormullion’s blog mentions Garry Trudeau’s use of an asterisk to represent George W. Bush. There was actually something of a fortunate coincidence that led to this. W.’s father, the first President Bush, was always depicted in Doonesbury as a dot, signifying his lack of substance and distinctness from Ronald Reagan. (He once, after telling a crowd that he had in fact, more than once told Reagan that he thought Reagan’s ideas were “dead wrong,” started to materialize as “his own man.” However, he immediately reverted to being a substanceless dot when he claimed that, unfortunately, all such discussions were strictly classified.) So it fit perfectly for the second President Bush to be portrayed similarly, with no substance—and now his position marked by an asterisk denoting the phoniness of his electoral victory.

    The fortunate coincidence reminded me of how, more than a decade earlier, President Reagan’s repeated claims not to remember the key events of the Iran-Contra Affair had given Trudeau the perfect springboard to launch the “Return to Reagan’s Brain” series of strips.

  6. That reminds me — I was annoyed by the post’s reference to “the cartoonist Doonesbury,” and I imagine Trudeau will be even more so if he ever sees it.

  7. J.W. Brewer says

    Obviously the controversy over Roger Maris in 1961 is a classic example of the deprecatory sense of “asterisk,” but now I’m curious if that’s where it originated or whether there are prior instances of using it literally or metaphorically to note that a claim was only dubiously true because of some apples-to-oranges problem in the definitions.

  8. @J.W. Brewer: I doubt that was where it originated, since the single-season home run record was never actually listed with an asterisk. Until the 1990s, there were always two records listed: Babe Ruth with 60 home runs for a 154-game season and Roger Maris with 61 home runs for a 162-game season. Every source of baseball stats seemed to faithfully reproduce the dual records, until Major League Baseball finally, belatedly decided to drop Ruth’s record from the books. (What I actually remember wondering when I first learned of the Ruth-Maris single-season home run controversy was: Why had they increased the number of games to 162 in the first place? Did people think 154 games was somehow too few?)

    In a similar vein, I remember that when Barry Bonds was on the verge of breaking the career home run record, it became common for people in the stands at Giants away games (especially in the right field stands, where Bonds tended to pull his home runs) to hold up signs with asterisks on them, denoting the tainting of his records by his steroid use.

    I also just checked the OED entry for asterisk, and it looks to have never been updated since 1885. However, it does include a 1645 attestation (“Set out as imperfect with three asteriscs.”) that suggests that the relevant sense: “The figure of a star (*) used in writing and printing:… (c) to distinguish words and phrases as conjectural, obscure, or bearing some other specified character…,” may be quite old.

  9. David Marjanović says

    “the cartoonist Doonesbury,”

    “Tim Apple”

  10. A columnist at the SF Chronicle once asked for submissions of (made up) clueless questions. One of them was, “So what other books did Harry Porter write?”

  11. Lars Mathiesen says

    The use of the asterisk for wildcard file name matching probably stems from Kleene’s use of a superscript star to denote closure of a set under string concatenation, part of the regular expression ‘language’. When regular expressions had to be in string form themselves, the ASCII / FIELDATA / EBCDIC asterisk (and question mark) were used. For wildcard matching, there is only one subexpression that you want to repeat, the match-anything, so that was cleverly elided and * used for the regular expression .*.

    The RT-11 monitor that I cut my sysadmin teeth on in 1979 had wildcard matching using *, but I think there was a lot of exchange between Digital and MIT in the 70s. It’s very likely that Dennis did it.

    (The use of a superscript star specifically for set closure is probably hard to trace, it had and has many different uses in mathematical notation).

  12. Stu Clayton says

    Dennis ?

  13. PlasticPaddy says

    Do you mean this Dennis?

  14. Lars Mathiesen says

    Dennis Ritchie, co-inventor of UNIX. Back when I got started, we almost felt like we knew Brian (Kernighan) and Dennis.

  15. John Cowan says

    There was a discussion of this on one of the Unix Heritage Society mailing lists, either tuhs@ or coff@. The * can be explained by Kleene’s notation, but not the ?. It looks to me like the combination first surfaced in either Multics or the DEC PDP-6 Monitor (the direct ancestor of TOPS-10), both form 1964, but it’s hard to be sure because the feature may not have been present from the beginning. I had thought that the Unix (aka Thompson) shell was the first to do its own wildcard expansion rather than foisting it off on the program being run, but that turned out not to quite be the case either.

  16. Lars Mathiesen says

    Kleene’s minimalistic syntax for regular languages only had alternation, you’d express the modern regex ? by alternation with the empty string.

    But qed for CTSS had regular expressions, it seems, so I assume Ken Thompson put that bit of syntax in then. If I had to pick an ASCII char to mean ‘maybe’ I’d probably make the same choice.

    The wildcard function of ? is not as closely related to the regex function as that of * is, but the actual length-one wildcard in regexes is . which would be confusing when wildcard-matching file names. (Even though filename parts separated with . were not important in Unix, I think people were used to them from other systems).

Speak Your Mind