Of sommeliers and statistics

Winemaker and statistician Robert Hodgson wanted to know the science behind why his wines sometimes won grand accolades, and why they faceplanted other times. So he applied rigorous statistics to the highly subjective sport of wine tasting, with fascinating results. From The Guardian:

“Only about 10% of judges are consistent and those judges who were consistent one year were ordinary the next year. Chance has a great deal to do with the awards that wines win.”

You can find the details of the entire study in the Journal of Wine Economics (of course there is such a thing) – Volume 3, Issue 02, Winter 2008, pp 105-113. Abstract here:

Wine judge performance at a major wine competition has been analyzed from 2005 to 2008 using replicate samples. Each panel of four expert judges received a flight of 30 wines imbedded with triplicate samples poured from the same bottle. Between 65 and 70 judges were tested each year. About 10 percent of the judges were able to replicate their score within a single medal group. Another 10 percent, on occasion, scored the same wine Bronze to Gold. Judges tend to be more consistent in what they don’t like than what they do. An analysis of variance covering every panel over the study period indicates only about half of the panels presented awards based solely on wine quality. (JEL Classification: Q13, Q19)

The whole Guardian article is a hoot, including snippets about how music influences perceptions of wine quality (Jimi Hendrix improves Cab Sav while Kylie Minogue adds a boost to Chardonnay). Unsurprising. My guess is that Van Morrison is the muse of Pinot Noir and Yo Yo Ma pairs well with Gewurtztraminer. The best anecdote from the article is at the end:

In 2007, Richard E Quandt, a Princeton economics professor, published a paper entitled On Wine Bullshit: Some New Software?” The study sought to describe the “unholy union” of “bullshit and bullshit artists who are impelled to comment on it”, in this case wine and wine critics. Quandt compiled a “vocabulary of wine descriptors” containing 123 terms from “angular” to “violets” via other nonsense descriptions such as “fireplace” and “tannins, fine-grained”.

Then, with the help of colleagues, he built an algorithm that generated wine reviews of hypothetical wines using his “vocabulary of bullshit”. For instance: “Château L’Ordure Pomerol, 2004. Fine minerality, dried apricots and cedar characterise this sage-laden wine bursting with black fruit and toasty oak.” He concluded that whether his reviews were “any more bullshit” than real ones was a “judgment call”. Sadly, he didn’t explore how long it would take a monkey to type a wine review.

Chaos theory and international relations

Foreign Policy is knocking it out of the ballpark this week! Fascinating post about the relationship between chaos theory and international relations. In essence, small perturbations initially can cause large disturbances subsequently, hence the analogy of a butterfly flapping its wings setting off a hurricane halfway across the globe.

Perhaps chaos theory could be most applicable to natural disaster research. Natural disasters can start small, affecting a few, but build to a momentous political maelstrom. The article quotes one of my favorite scholars, Kalev Leetaru:

Today, most theorists have more modest goals. Chaotic systems are extremely difficult to predict in the long run, but they’re also not entirely random – as Lorenz observed – and with enough detailed information, patterns emerge allowing short-term predictions to be made, though always with a degree of uncertainty. As Kalev Leetaru told me recently discussing the GDELT events database, “Most datasets that measure human society, when you plot them out, don’t follow these nice beautiful curves,” he says. They’re very noisy because they reflect reality. So mathematical techniques now let us peer through that to say, what are the underlying patterns we see in all this.”

Linguistics and The Economist

I just love it when The Economist posts about linguistics. I mean, could you name your favorite era in linguistic history?

IF FORCED to pick my favourite part of the history of English, I’d be torn. There are so many to choose from. Would I pick the Great Vowel Shift, the mid-millennium change in pronunciation that largely explains English’s inconsistent spelling? Perhaps I’d turn to colonial times, when English vocabulary ballooned. I do like Noah Webster’s attempts to change American English spelling in the name of efficiency, too.

From this post, we can learn new words like diglossic:

During this period, England’s society was diglossic: one community, two language sets with distinct social spheres.

It’s about vocabulary selection and register, or the context and setting which informs our decisions about which words to use.

Although English is no longer in a diglossic relationship with another language, the Norman-era diglossia remains reflected in the way we choose and mix vocabulary. In informal chat, for example, we might go on to ask something, but in formal speech we’d proceed to inquire. There are hundreds of such pairs: match/correspond, mean/intend, see/perceive, speak/converse. Most of us choose one or the other without even thinking about the history behind the split. Germanic words are often described as earthier, simpler, and friendlier. Latinate vocabulary, on the other hand, is lofty and elite. It’s amazing that nine hundred years later, the social and political structure of 12th-century England still affects how we think about and use English.

It’s also about deeply embedded linguistic multiculturalism, and how peoples integrate languages through conquest and migration. It is fascinating that these histories are obscured by hundreds of years of usage, taken for granted as interchangeable options dependent on the situation. When we evaluate the way in which someone speaks, at least in diglossic or formerly diglossic languages, we inevitably become linguistic archeologists.

Many languages have “high” and “low” layers of vocabulary. But in most other languages, the two sets are drawn from the same source. By contrast, contact between Old English and French, Dravidian languages and Sanskrit, Japanese and Chinese, Persian and Arabic, and other pairings around the world have created fascinatingly hybrid languages. These mixed lexicons are, for linguistic and social historians, akin to the layers of fossils that teach paleontologists and archaeologists so much about eras gone by.

Disease and development

Foreign Policy magazine asks the following question:

How did Europe go from a global backwater around 1400 — defined by political fragmentation, poverty, and widespread illiteracy — to the most prosperous region the world had ever known by the dawn of the 18th century?

And they get a strange answer: the bubonic plague.

(t)he population shock of the Black Death was so dramatic that it caused a permanent increase in incomes — an estimated 30 percent in Western Europe between 1500 and 1700. The authors note that China, in terms of farm wages, urbanization, and economic output, was pretty much on par with Europe when the Black Death hit. Had Europe’s population continued to grow unchecked during this period, Voigtländer says, it would have been at about the development level of China in 1700. Instead, Europe by 1700 had raced far ahead, setting the stage for its next breakthrough: the innovation and rapid increase in trade that began during the Industrial Revolution decades later.

Research puzzles

Reading the Little House on the Prairie series was one of the literary highlights of my childhood. The stories about Laura Ingalls Wilder and her life on America’s frontier were fascinating and compelling. Researchers have recently begun to explore one of the puzzles contained in this book: the case of Laura’s sister Mary’s blindness.

Researchers have deduced that Mary had viral meningoencephalitis, not scarlet fever as Laura wrote in her memoirs.

It may not be the biggest bombshell to hit the medical world, but to “Little House” fans, the question remains: why did Wilder change her sister’s illness to scarlet fever? The study authors believe it could be because Wilder and her editors thought scarlet fever would be more relatable to her readers. Scarlet fever is mentioned in other books from the period, including “Little Women” and “Frankenstein.”

This goes to show that puzzles can arise from anywhere, and that investigating them can yield new insights, and new questions.

Eat chocolate, win Nobel prize

I should be well on my way.

“When you correlate the two – the chocolate consumption with the number of Nobel prize laureates per capita – there is an incredibly close relationship,” (the study’s author Franz Messerli) says. “This correlation has a ‘P value’ of 0.0001. This means there is a less than one-in-10,000 probability that this correlation is simply down to chance.” (h/t BBC)