Re: how to sort by stroke (not radical/stroke)

From: Andrew C. West (
Date: Mon May 19 2003 - 10:15:29 EDT

  • Next message: "Re: Decimal separator with more than one character?"

    On Sun, 18 May 2003 17:55:15 -0700, "Allen Haaheim" wrote:

    > Andrew, Thank you for your thorough summary. I failed to make clear that I
    > was talking about the implicit grouping of characters *below* the explicit
    > subsort level in the _Ci hai_ index, and was wrong to say that this implicit
    > subsort was by radical. Rather, they are subsorted by stroke category (I
    > think?), but definitely not radical, although it looks like radical because
    > same-radical characters tend to be lumped together due to stroke category
    > sorting method. At any rate, there still appears to be some method of
    > implicit subsort below the explicit first-two-strokes (qimao) subsort in
    > square brackets, which seems to yield arbitrary groups of smaller size
    > because of the three levels involved.

    Yes, I suppose that the implicit third level sort is by stroke category of the
    third plus strokes. Which means that Marco was right that you would need an
    ideographic database that breaks down characters into all their component
    strokes in order to sort by stroke correctly.

    > In my previous emails as well as here, I was thinking of dictionaries that
    > deal at least in part with traditional (Classical) Chinese, published
    > anywhere, so I included dictionaries such as Karlgren and Mathews, but
    > excluded dictionaries strictly of modern Chinese. I think this generally
    > raises the prevalence of radical/stroke indexes and of course lowers the
    > number of pinyin-based indexes. Below are all the books I used.

    Probably the opposite, as modern Chinese dictionaries such as _Xiandai Hanyu
    Cidian_ and _Xinhua Zidian_ are internally ordered by pinyin, and so have
    Radical/Stroke indexes, but have no need for pinyin indexes, whereas more
    "classical" dictionaries such as _Ci Yuan_ and _Ci Hai_ tend to be ordered
    internally by Radical/Stroke, and so have no need for a Radical/Stroke index,
    but may have a pinyin or other phonetic index.

    > I'm also curious about the _Peiwen yunfu_ and _Kangxi zidian_, which I don't
    > have at home but are oft-consulted works.

    _Kangxi Zidian_ is ordered internally by Radical/Stroke (the famous Kangxi
    radicals), but the original edition does not have any index other than a short
    Stroke Count index (subordered by Radical if I recall correctly) of hard to find
    ideographs. The modern Zhonghua Shuju reprints of _Kangxi Zidian_ do not add any
    indexing method, but maybe some other modern editions that I haven't used do.

    _Peiwen Yunfu_ is not a dictionary as such, but [as I recall, not having had any
    cause to use it since I was an undergraduate many years ago] rather a
    concordance of literary words and phrases ordered by rhyme (see [no relation]).

    > You must have a different edition than my _Ci Yuan_ (Shangwu yinshuguan,
    > [xiudingben] hedingben 1-4, 1997, 1 vol.). In mine the primary indexes are
    > Pinyin and Four-corners, although there is a short "Hard-to-Find Characters"
    > index by stroke count/stroke category. The internal ordering is by
    > radical/stroke and is all divided into 12 parts by Earthly Branch, and
    > printed as running heads (a handy feature).

    Exactly the same edition as mine. The index I was referring to is indeed the
    short "Hard-to-Find Characters" index at the front of book.

    > According to your survey, I also stand corrected on radical/stroke as the
    > "most commonly found" index, though my personal sampling of about ten
    > dictionaries still is mostly radical/stroke. And it is good to emphasize, as
    > we've both done, that usually a stroke count index is primary only because
    > the dictionary in question is already internally ordered by radical/stroke.

    100% agreement on that one.

    > Please point out any more errors I might have made.

    Indeed, is not pointing out other people's errors the main raison d'etre for
    this list ;)


    This archive was generated by hypermail 2.1.5 : Mon May 19 2003 - 11:07:59 EDT