23.7 aztranslate limits


AI remains a dream for now. Whilst our current translation tools do a pretty fine job translating from one language to another, it does so, as best we understand, with limited or no understanding of the text it is translating. The limits command shares a short article on the limits of automatic translation.

Douglas Hofstadter, a professor of cognitive science and comparative literature at Indiana University at Bloomington and author of the highly recommended book GΓΆdel, Escher, Bach: an Eternal Golden Braid, highlights in a January 2018 article in The Atlantic the limitations of automated language translation. To paraphrase, the translators do not have any deep understanding of the text but have developed a shallower mechanical process to do a decent job for simple communications.

Below we illustrate with one of Hofstadter’s examples which you can replicate with the limits command. See the original article for details:


$ ml limits aztranslate


*** Consider this sample text:

In their house, everything comes in pairs. There's his car and her
car, his towels and her towels, and his library and hers.

*** The French translation is:

Dans leur maison, tout se passe par paires. Il y a sa voiture, sa
voiture, ses serviettes, ses serviettes, sa bibliothèque et la sienne.

*** Translating back to English demonstrates a shallow understanding:

In their House, everything happens in pairs. There's his car, his car,
his towels, his towels, his library and hers.

Your donation will support ongoing availability and give you access to the PDF version of this book. Desktop Survival Guides include Data Science, GNU/Linux, and MLHub. Books available on Amazon include Data Mining with Rattle and Essentials of Data Science. Popular open source software includes rattle, wajig, and mlhub. Hosted by Togaware, a pioneer of free and open source software since 1984. Copyright © 1995-2022 Graham.Williams@togaware.com Creative Commons Attribution-ShareAlike 4.0