I keep meaning to post versions of the various talks I've given (the ones that are not videos) but I haven't yet, for myriad reasons … including trying to get actual work done. But the Ignite talk I gave was pretty short (five minutes!), so I thought I'd try to put it up here.

Slide 1: I always include a definition of 'lexicographer' when I talk; keeps people from elbowing their neighbor and asking 'what did she say she did again'?

Slide 2: Because this was a tech talk, I also pointed out that I was a geek. (I wore this skirt, by the way, for visual reinforcement of the concept.)

Slide 3: We all think of dictionaries as very concrete, solid objects. (You wouldn't want to drop one on your foot, would you?) But actually …

Slide 4: Dictionaries might be solid, but their innards are really collections of ABSTRACTIONS. Definitions are made by taking a lot of specific datapoints (uses of a word in context) …

Slide 5: … and averaging them out to a more general meaning.

Slide 6: This may seem really obvious to YOU … [note: this is my favorite Flickr image EVER]

Slide 7: But many people think that lexicographers just "decide" what a word means. Nope! That would be really hard work … it's easier (and more accurate) to look at examples of use.

Slide 8: What can you tell about this word [pirgate] from these examples? Well, I bet you know it's a verb, that you can do it TO something, and that it's something you might not want to do. This is all information you know implicitly because you know how English works.

Slide 9: How about now? Now you know this use is a noun, and it's a kind of person — the kind of person you don't want to be, probably.

Slide 10: The truth is that "meaning" is created by lots of little points of data, in the same way that persistence-of-vision effects are created by lots of little points of light. A diffuse set of data can look pretty solid if it moves fast enough …

Slide 11: But you really do have to have ENOUGH data for this persistence-of-vision effect to work. If I say a guy is wearing a tux and holding a martini, is he a waiter, or is he James Bond? You don't have enough data to tell.

Slide 12: So lexicographers in their labs try to distill all those usage data points into high-octane liquid definitions.

Slide 13: Now, if you want to brew your own, and want a lot of examples of use, there are more places than ever to try to find them …

Slide 14: Although it doesn't work for every word … (especially not 'pirgate', since I made it up).

Slide 16: And, of course, the examples you DO find might not be helpful.

Slide 16: The big question, though, isn't really WHERE to find enough examples — that's pretty straightforward. The big question is: if we agree that dictionary definitions are abstractions of meaning, is there a better way to represent those abstractions than this:

Slide 17: … the boring old print dictionary?

Slide 18: Could we show relationships between examples in a less-linear way?

Slide 19: Could we convey those abstractions in a more powerful way?

Slide 20: That's the problem I'm working on every day … [go visit my blog, yadda yadda, big plug for Creative Commons and the awesome nice sharing people on Flickr.]
(All my presentations are Creative Commons-sharealike, by the way, so if you ever want to remix me into a rap song or create a dictionary-talk novel or whatever, go right on ahead. If you see me speak somewhere and want a copy of my slide deck, just email and ask.]
And that's what I talked about at Ignite last Thursday. More or less. I didn't make any notes, so this is from what I remember of what I said off the top of my head!

Labels: speaking, talks