Thursday, December 16, 2010

which color synonym should I use?

Recently we wrote about magenta and fuchsia being synonyms, and on Water Day we wrote about the difference between cyan, aqua, and (for the horizontal readers of our blog using the etymology tag) turquoise, (see also here, since we are still in 2010). The question we left open was how to choose between color synonyms. As color scientists, we do not want to come across as being too geeky or おたく

From today, there is a tool to answer this question.

In the old days of color science, in psychophysics we dealt with about 15 subjects (logarithmically: 1). Then with the multilingual color naming experiment, HP Labs ratched this up to thousands (logarithmically: 3). Today, Google Labs is taking this to the next level: millions (logarithmically: 6), or 5,195,769 observers if you need an exact number.

When the color naming experiment allowed you to find color names and their synonyms, the new corpus allows you to decide which synonyms to use. Let us get back to magenta and fuchsia. You are at Silicon Valley cocktail party, want to talk about your daughter painting her room in hot lips color, but do not want to come across as a geek. Should you call it magenta or fuchsia?

First learn our latest tag: culturomics. While our technological advantage was crowd-sourcing, the next level is culturomics, where instead of harvesting an audience, all books are interrogated. Moreover, while crowd-sourcing was ephemeral, culturomics introduces time as a new parameter. The details are in the upcoming Science magazine paper Quantitative Analysis of Culture Using Millions of Digitized Books, by Jean-Baptiste Michel1, Yuan Kui Shen, Aviva P. Aiden, Adrian Veres, Matthew K. Gray, The Google Books Team, Joseph P. Pickett, Dale Hoiberg, Dan Clancy, Peter Norvig, Jon Orwant, Steven Pinker, Martin A. Nowak and Erez Lieberman Aiden.

Essentially, Google Labs today is making 5,195,769 of the 15 million books they have digitized available for research. To put the number in perspective, these are about 4% of all books ever published.

So, is it magenta or fuchsia? [click on the graph to view it at full size]

use of magenta vs. fuchsia

In harmony with what we wrote, fuchsia has been around since 1730, while magenta emerged 1851 with the battle in Magenta. While the first use in books was probably regarding Magenta, uses after London's 1860 red magenta definitively refer to the color. The above graph teaches us, that when talking to non-scientists we should use magenta rather than fuchsia, even when the former is currently dipping.

This is good. After all, the color naming experiment teaches us that nobody knows how to spell fuchsia anyway.

Now, let us look at cyan. [Click on the graph to view it at full size]

use of cyan

The corpus queried was English books, specifically, Italian books were excluded. Although the Latin spelling aqua has always dominated, up to 1870 the Italian spelling acqua was sometimes used. At the cocktail party, however, you should call it turquoise.

Although this is a quantum leap, it does not yet solve the hot lips problem, because there is not enough data to isolate the context of color. For that we will have to progress beyond books, but Google does not yet know how to structure its corpus, as we wrote a while back. Will it be Bing?

No comments:

Post a Comment