A sofa, a book, a knife, floating in a pool

« previous post | next post »

Helen DeWitt at Paperpools describes an interesting classroom experiment in techniques for (adult) learning of grammatical gender ("Winnie der Pooh", 5/15/2008)

18 students were tested on an artificial-language task, where gender was marked by choice among different articles, as in German:

[They] were given a list of 20 words, each of which had been assigned one of three articles invented for the occasion (fif, led, had). They were given three minutes to memorise the articles; asked to chat among themselves for two minutes; then given a test on the articles.

The number of correct replies reported was:

2 6 6 7 8 8 9 9 13 14 14 15 15 16 16 17 18 19

In this first test, no particular technique was prescribed, but most students tried to memorize the word list for each article type, or to memorize two of the three lists and assign the third by default.

The student who got 19 out of 20 "said she constructed a story using images from one group of nouns, another story for another." Techniques like that one, based on associative imagery, seem to work better on such tasks than list-memorization does. Helen observes that this is

… essentially that of the art of memory as practised in antiquity, one I've heard of but never bothered to try: one assigns an image to the thing to be remembered. So one might assign the image of water to 'das' and then, to remember the gender of a neuter noun, assign a mental image to the noun and associate this with water. […] In one's mind, then, one might have a sofa, a book, a knife, floating in a pool.

When the students try the visualization technique on another list of "20 nouns and 3 invented articles", the results are:

(pairing results with original results):

18 20 20 20 20 18 20 19 19 18 17 20 20 19 20 20 19 19

This reminds me, a little, of Lera Boroditsky's work on "Sex, Syntax and Semantics" (discussed in "Sapir/Whorf: Sex (pro) and space (anti)", 11/19/2003), though the associations that Lera demonstrated in native speakers are apparently more complex and diffuse, and in any case result from natural experience rather than from conscious imagery. (Also, native speakers are sometimes less clear about grammatical gender than you might expect: see Heidi Harley's post "You say feminine, I say masculine, let's call the whole thing off", 2/28/2008.; as well as "Clarifying status in Wolof by fake disfluency", 5/20/2004.)

More to the point, though, I wonder whether imagery techniques scale well in learning morphological categories for hundreds or thousands of words. And I also wonder whether such techniques are commonly used in language teaching. I've certainly never experienced them in a language class, and apparently the application was new to Helen as well.


  1. panne said,

    May 16, 2008 @ 9:41 am

    I have never experienced such a technique used in language teaching, but I hope it will become common (and that it works) – I remember being pretty good at German grammar in school, but failing really hard at remebering genders, even though we have the same three genders in my native language. It's one of the great things about English; no genders!

    As a sidenote, I have often thought that there must be some logic behind gender assignment on nouns – of there wasn't, native speakers wouldn't agree to such a big extent on the gender of novel nouns that enter the language.

  2. David Lurie said,

    May 16, 2008 @ 10:30 am

    Similar techniques are used in the study of Chinese characters, presumably in Chinese-language contexts, and definitely in Japanese (i.e., kanji). The best-known presentation of this approach to character memorization is that of James Heisig, in a 1977 book called "Remembering the Kanji" (reissued, with a second volume sequel, in 1987) but I think that image-based character mnemonics have probably been 'invented' a number of times. Marshall Unger discusses them in the context of stage magicians' memory tricks in "Ideogram: Chinese Characters and the Myth of Disembodied Meaning" (2004).

  3. Bruce Rusk said,

    May 16, 2008 @ 10:36 am

    Some Jesuit missionaries to late-Ming China used the classical technique as an aid in their study of Chinese. And one of them taught it to Chinese students to help with the memorization of canonical texts.

    See Jonathan Spence, The Memory Palace of Matteo Ricci (NY: Viking Penguin, 1984).

  4. Coby Lubliner said,

    May 16, 2008 @ 11:58 am

    It's hard for me to see what this experiment has to do with the actual remembering of grammatical gender in real languages. My first primary ("native") language was Polish; it was succeeded by Hebrew, and then German. All three have grammatical gender, but in Polish and Hebrew gender is usually indicated by the word ending (with more exceptions in Hebrew than in Polish) while the definite article is either nonexistent (in Polish) or ungendered (in Hebrew). Only German conforms with the "language" of the experiment, in that the form of the noun gives no indication of its gender (unless it has a suffix such as -in or -heit). And in fact I have more trouble remembering gender in German than in the other two.

  5. Sili said,

    May 16, 2008 @ 7:31 pm

    Years ago I saw a similar suggestion for learning secondary languages. Only it focused on visualising a city or neighbourhood one was familiar with, then divide it into as many (natural) sections as necessary and then distributing the nouns in these sections according to gender.

    It sounds good in principle, but I'm singularly lacking in imagination, so I've never been able to engage in the exercise. I can't help but find it ridiculous – and essentially as unnatural and as much work as rote learning. Which I also hate.

    Obviously I'm only sesquilingual as a result.

  6. Kassy said,

    May 16, 2008 @ 7:32 pm

    I had a professor who learned Old English from a book which, rather than using images, ascribed a colour to each gender. I don't know if that would be considered to be the same technique. Unfortunately I don't remember the name of the book, and I believe he mentioned that later editions discontinued the colour coding.

  7. Ran Ari-Gur said,

    May 16, 2008 @ 9:28 pm

    @panne: There might be a pattern for assigning genders to neologisms and new loanwords, but there's no guarantee that the pattern applies to all or even most existing nouns.

  8. Stephen Jones said,

    May 16, 2008 @ 10:42 pm

    A friend of mine who worked at the Complutense in Madrid some years back gave a set of made-up Spanish words to his students (none of which had the give away o/a endings) and asked them to assign the gender. There was a very high correlation between the different students gender assignations.

  9. gordonoz said,

    May 17, 2008 @ 7:19 am

    The English novelist Anthony Burgess, somewhere in his prolific writing about language, says that he is surprised that nobody has noticed the reason why the French (and others) use female and male genders for nouns. On the whole (though not 100%) the male gender nouns project and can be inserted, whereas the female ones have things put in them or on them, echoing the penis and vagina respectively. So with la table you can put things on it and in its area, with la fenetre and la porte you can go through them, le crayon is pointy, as is le nez, but la bouche receives. There are many, many exceptions, but a sufficient number of nouns accord with this theory to make it useful as a starting point.

  10. Randy Alexander said,

    May 17, 2008 @ 7:44 am

    I wouldn't say such techniques are so common, but I have used them from time to time. When memorizing Japanese vocabulary, I made color-coded flashcards; verbs were on yellow cards if they were ichi-dan, and on pink cards if they were go-dan (the two conjugations).

    I have taught the memory palace technique to students in China when they wanted to memorize lots of words for vocabulary tests. It's a very powerful tool for memorization, but memorizing huge amounts of words without really knowing how to use them isn't so useful, so I don't teach this method so often.

    I teach English to Chinese elementary school students, and I have noticed that when I teach words with irregular spelling, the kids can remember the pronunciation of the those words much more easily if I use a different font on the flashcards for those words than for the regularly spelled words. Recently I tried using the same font for all words, and this ability fell apart.

  11. Aaron Davies said,

    May 17, 2008 @ 8:30 pm

    All through high school French, I remember the gender of the first forty or so nouns I'd learned by recalling which page of a two-page spread in our textbook they'd been on–masculine were on the left, feminine on the right.

  12. Jason Eisner said,

    May 17, 2008 @ 11:21 pm

    Mark asks: "I also wonder whether such techniques are commonly used in language teaching. I've certainly never experienced them in a language class …"

    As a kid, I read "The Memory Book" by Harry Lorayne and Jerry Lucas. They recommended devising ridiculous images as mnemonics for foreign-language vocabulary. For example, to remember that grapefruit in French is pamplemousse, visualize a pimply moose whose pimples are actually grapefruits. Not only do I remember this image of theirs many years later, but I found the technique useful from time to time when studying French.

    I don't remember whether they said anything about memorizing gender or other morphological categories — but they may have.

    In general, the book made heavy use of associative images, explicitly building on the methods of antiquity that Helen mentions. Much of the book involves systems that help you to turn various kinds of data (e.g., digit sequences) into memorable images or sequences of images.

    Images presumably aren't the only solution. Indeed, I have always remembered phone numbers by looking for mathematical patterns among the digits or groups of digits (my wife does this too), or sometimes visual patterns on the phone keypad, or musical patterns by associating the digits with notes. Regardless of the medium, remembering or half-remembering one or more mnemonics ought to provide you with some useful redundancy: you have an overconstrained representation of the data. Furthermore, you probably get some advantage just from having dwelled on the data long enough to try to craft your mnemonic (even if you fail to generate and remember one).

  13. Jimmy said,

    May 19, 2008 @ 8:19 am

    Don't German learners use a mish-mosh of "rules" to help them learn articles? I know I did. They aren't 100% accurate, but they'll get you pretty far:

    * ends in "e" – usually "die"
    * one syllable – usually der/das (at least you're down to a 50/50 chance)
    * foreign words – often das
    * ends in "er" – likely "der"

    After you have enough experience, some articles just sound right, which I guess is what it's like for a German learning German. I'm not sure is using an analogy would help a learner acquire this skill.

  14. leonardo o higgins said,

    June 2, 2008 @ 11:00 am

    Obviously Sili, you are brilliant.

RSS feed for comments on this post