A long time ago I read chunks of a book by Douglas Hofstadter titled Metamagical Themas which has a section on how we manage to see the same letters represented in what are, in fact, many different ways. He uses the example of Chinese characters as an illuminating example because, while to us Chinese characters in different fonts often look startlingly different, to a native Chinese reader they’re instantaneously recognizable as the same character. But it’s the same with fonts. It seems that there are a wide range of things that are highly variable in a written character’s (read: “letter’s”) makeup that we can tolerate being altered without the character losing any of its legibility.

I’ve long thought that the same of spoken languages. While many of my friends had trouble understanding my parents’ speech, I never had a problem hearing through their accents, but what always confounded me was how I was heading my mother say, “Don’ say dat dammet” and I was understanding this to mean the words, “Don’t say that, dammit!” I realized that my own tolerances must have been different from those of my friends, but what I also realized was that English having so many different accents, many of them instantly (or quickly) penetrable to the native listener, meant that human tolerances for spoken pronunciation must be high in a number of areas. A lot of variations must be possible before a spoken word becomes incomprehensible.

(This has probably been tested in the laboratory. Perhaps Hofstadter has even written about it. But what’s interesting is that all my searches lead me to results of studies in which people are trying to get computers to recognize speech, or to simulate it conmvincingly. However, I’m sure studies on humans abound… I’ve found abstracts online that suggest it.)

I’m not sure if my observations reflect the idiosyncracies of people I’ve known and taught or whether I’m right, but I have developed a sense, over the last few years, that each language, when spoken by native speakers, has a different set of tolerances for variation which speakers become attuned to. For example, it seems to me that while stresses in the wrong place in English seem to be amusing and funny, we can often still piece together what someone is saying to us. Perhaps it’s just a lack of exposure to people speaking this way in Korean (the way “foreign accents” are run of the mill in English), but I find that if one does this with older Koreans, they usually simply cannot figure out what the hell you are saying to them.

Meanwhile, it’s funny because, for a language with so many precise variations on consonantal sounds (b, Bb, p all being related, g, Gg, k being similarly related, and so on), I find that native speaking Koreans actually have a fairly high tolerance for weirdly pronounced consonants. Bs becoming Bbs and Ps seem not to matter much—which is probably part of that filling-in-the-blanks thing that’s so common in much of human cognition.

Anyway, because of this, I think, today brought me one priceless moment where my students were doing an exercise in which they had to discuss their childhood favorites. “What was your favorite game? What was your favorite toy?”

One group called me over to ask me the English word for 만�? (mandu). When I told them it was “dumplings,” and reassured one cautious fellow that yes, this is a noun, not a verb (he saw the “-ing” and began to worry), one student answered the question, “My favorite food was crap dumplings!”

I’m trying not to laugh as I look at him and try to figure out what the hell he’s saying. After a few moments, when one of the guys explains in Korea that he means 게만�? I laugh and explained how he was pronouncing the word wrong, and why he should be careful with that particular word. You should have seen the riot of laughter that exploded in his little group when I explained what “crap” means in English. I think the poor fellow actually visualized dumplings filled with human excrement, he looked so shocked. “No no no no no! CRAB dumplings! I like CRAB dumplings!”