NNS * NNS = 144

Abstract: The present dictionary interface is the same kind of information display as a multiplication table; current online dictionaries are putting big, searchable lexical multiplication tables on the web. But today’s dictionary data allows much more. Who/what will spark the dictionary-display revolution?

Body: Consider the multiplication table: a mute, ignorant record of calculations someone has done somewhere else. If it only goes up to 9*9 it can give you no indication of what 9*10 is. You can use your own mind to add or extrapolate, of course, but the paper can’t: WYSIAYG. A pocket calculator supersedes the multiplication table and offers at least the arithmetic operations. The pocket calculator meets a user who is in the same state as a dictionary user: “I know my question; this will tell me the answer”. But instead of simply showing all the answers all at once and letting the user grep optically for the right answer, the pocket calculator asks the user to enter exactly what she is calculating, and in return she gets only the result that she requested: YAFIYGI.

Any computer that can get on the web obviously supersedes even the pocket calculator. People don’t put multiplication tables on the web to help you find the answer to simple multiplication problem. You do find multiplication tables on the web (some of them are very cool), but they’re a tool to help you learn mathematical concepts, not to help you find out 6*7.

Both multiplication tables and lexical lists were inscribed on clay tablets in ancient Sumer 5000 years ago, and both have been made fairly continuously since. The sophistication of quantitative display has progressed at a pace matching general technological/scientific developments; but dictionaries, in their own transition from clay to screen, haven’t even reached the pocket calculator stage. A dictionary is great for what it’s been doing for centuries, but the surrounding science has progressed far beyond what you will see in any existing form.

A paper dictionary is the same sort of thing as a multiplication table: the output of a ton of research and thinking, fixed immutably for linear consultation. Web dictionaries, in turn, have put the paper layout on the screen and made only slight improvements on the paper version. You may jump quickly to a headword, maybe search for a word within definitions, and maybe even search for a word in the definition of a lemma in a particular part of speech. But still, however sophisticated your search, the output is the same: an entry as it would appear on the page. Garbage in, entry out; pearls in, entry out.

Yet a mature, modern dictionary, in its native electrons, has all kinds of information that doesn’t fit on the page or into any current ideas of what a dictionary can tell you. Somehow, sometime soon, someone will present dictionary data not simply as multiplication, but as the full range of arithmetic and other mathematical functions: +-*/^!% -> lexical, semantic, historical, phonetic, syntactic, collocational, sociolinguistic, geographical, etc. etc.; and all-importantly, it will be context-sensitive. Yes, if it ain’t broke, don’t fix it; but there’s nothing broke about the multiplication table: it’s just not functional enough for daily use as an essential mathematical reference. Current dictionaries work fine for what they do, but a more sophisticated lexical interface will allow people to ask more sophisticated lexical questions.
I want my computer’s dictionary to be a lexical Bloomberg terminal that reads my mind. With the industry’s most advanced dataset at my fingertips, I can wrap my mind around the contextual part, but I haven’t figured out how the new ink-liberated display will look. But I’m thinking about it, and folks like these are rich guano for my brain. So much of the computational lexicographical work I do is quantitative at every stage except the final output: where and how could, or should, the numbers come in to the final data? Numbers are part of a lexicographer’s decision-making process, and part of the amateur Google corpus ad-hoc definer’s process; but not part of dictionaries as we see them today. But tomorrow is coming, and soon.

Leave a Reply

Your email address will not be published. Required fields are marked *