Book Review: Jurafsky on The Language of Food (with implications for wine)

Dan Jurafsky, The Language of Food: A Linguist Reads the Menu. W.W. Norton & Company, 2014.

Dan Jurafsky is a Stanford University computational linguist who is fascinated by the way we talk about food. He’s written this provocative book that tells a series of stories that mainly trace the way that the language of food changes over time and how this is related to global cultural and economic exchange.

Turkey, turkey, sushi and ketchup

One story, for example, explain how the turkey, which is native to Southern Mexico, came to be named for the country of Turkey in England and here in the U.S. while acquiring names associated with India in France and elsewhere.

Another chapter traces the origin of ketchup (or sometimes catsup) back to China and then around the world to the familiar red Heinz bottle buried in the back of your refrigerator.  Great stuff if you are interested in food, globalization, history or language.

Unsurprisingly, I was hoping to read about wine in this book and I did both directly (the origins of the social custom of the  “toast” in spiced toasted bread that was dipped in wine to improve its flavor) and indirectly (vinegar, wine’s close relative, as the unlikely global connection between sushi, the food of Japan, English fish and chips, and a host of other delicacies).

Menu Language and Economics

The most thought-provoking ideas about wine actually came from the chapter of the language of restaurant menus. Jurafsky and his colleagues were able to amass a considerable database of restaurant menus for digital analysis. The data is both broad (there are hundreds of thousands of menus on the web) and deep (the New York Public Library has a historical collection of 10,000 menus dating back to 1843). So it is possible to analyze both how the language of the menu has changed over time and how if has evolved differently for different types of restaurants.

The menus don’t just describe food, they also list prices, which makes them interesting to an economist. What types of words are most associated with higher and lower menu prices? The researchers adopted the necessary control procedures (so that they were comparing apples with apples) and here’s what they found.

Language varied by the type of restaurant. Inexpensive restaurants tended to talk about the choice they provide their customers, which I suppose makes sense. Diners understand that inexpensive cafes often offer long menus and so standardized products are the norm. Choices (eggs as you like them, for example, or a dozen different flavors of dipping sauces and salad dressings) are a way to diminish the industrial quality of the experience.

More expensive fine dining restaurants don’t stress choice and, in fact, sometimes deny choice by having chef-chosen set tasting menus (sometimes even “blind” set menus where the list of dishes is not revealed when the diner maker her order).  Instead of stressing choice, the menus focus on the particular characteristics of the food, the origins of ingredients, and the nature of the preparation. Language literally “counts” here. There is a positive correlation between average word length in the menu descriptions and the price of the item. Each one letter increase in the average word length is worth about 18 cents!

You Said a Mouthfull!

Longer words and more specific terms are associated with higher menu prices. This is correlation not causation, of course, so it isn’t necessarily true that you and I see longer words or detailed descriptions and pull out our credit cards, but it could be true that these characteristics help us justify a higher price because they seem to signal to us higher quality. If wine drinkers have trouble avoiding the assumption that higher price signals higher quality, it’s not impossible that high-end diners respond to sophisticated word play.

Interesting finding: there seems to be a “spicy” or “exotic” tax. Menu items that include these words typically carry a higher average price than similar products without this designation.

So chicken you order at Wendy’s is about choice (do you prefer grilled or fried?) and at a fine dining restaurant it is about specific qualities and longer words (spicy Palliard of vegetarian-fed Draper Valley Farm chicken with exotic accompaniments), what about casual dining restaurants that occupy a vast middle ground?

Casual dining restaurants like TGIFriday’s, Applebee’s and Ruby Tuesday have a menu language all their own, stressing how the food tastes or is prepared, but in very broad general terms. Chicken is tender. Steak is juicy, Crab is … real (because you might think it could be fake crab). Interestingly, terms of this type are all associated with lower prices! It isn’t that you are thinking that tender chicken is worth less, it’s just that a higher quality place wouldn’t need to tell you that the steak is juicy or the crab is real, but a less expensive restaurant would.

Jurafsky calls these “filler words” and they are the types of things you say when you have to say something but there isn’t anything better or more specific to say. The more filler words on the menu, it seems, the less distinctive the actual food items and the lower the price. Interesting, isn’t it?

The Language of Wine

So what does this have to do with wine? Well, it seems to me that the same sort of research could be done on the language of wine, both how it has changed over the years and also the way that certain types of terms are associated with different categories of wine. I’m going to start paying more attention to wine advertising, wine label text and wine reviewer descriptions.

Some off the cuff observations are inevitable. If box wines are the vinous equivalent of fast food, then it makes sense that they would talk about convenience the same way that inexpensive restaurants talk about choice. No sense dwelling on low cost since that’s obvious. It’s the way that the product can be made to accommodate your desires that matters.

Very detailed descriptions reign at the top of the wine food chain just as they do for restaurant food — and this is perhaps because indicators of place and craftsmanship are now almost universally seen as indicators of quality and authenticity. It’s not just food or wine, it is pretty much everything that consumers look for.

Maybe this is why so many AVAs have been created in the U.S. — having an AVA associated with your wine is important to its credibility even if the consumer doesn’t really know what or where it is (and even if it doesn’t really mean very much in terms of quality or style of wine). Sue and I enjoyed a nice Vioginier recently that got my attention because it came from the Clarksburg AVA. Clarksburg? You don’t see that every day. Interesting! Gotta try it (and it was indeed interesting).

I’ve inserted an image of the back label of that Viognier above so that you can read the text. Pretty upscale messaging, don’t you think? Even the average word length is impressive.

What about the wine equivalent of casual dining — branded wines selling in the $8-$12 range?  A quick look at some labels suggests that the descriptors are just as vague (“juicy,” “delicious”) as on the casual dining menus and they probably fill the same function — fillers to provide a little (very little in some cases) textual weight when more specific terms don’t apply. Sometimes, I have noted, the labels can have very little to do with wine at all — simple but perhaps effective filler.

Robert Louis Stevenson said that wine is bottled poetry, but I think there is even more to it than that! The language of food and the language of wine seem to have something in common. Food for thought for wine marketers and consumers both!

4 responses

  1. Great post Mike! I’m glad you enjoyed our Viognier, thanks for the plug! Wilson Vineyards in Clarksburg has also put together a nice Chenin-Viognier, Petite Sirah, Sauvignon Blanc, and Chardonnay. Most Importantly, thank you for highlighting Clarksburg as an AVA. We have a lot of great growers and wineries in the area that have made great strides in the industry. Come on out and I’ll give you a personal tour.

    Phil

  2. Hi Mike, I’m a Psychology PhD Candidate at Harvard – my work focuses on how the brain allows us to understand other people – but in my spare time I do a bit of data science on various topics, including wine.

    I completely agree with you that we can learn interesting things from the language surrounding wine! In fact, I’ve done a bit of work in that line already: you might be interested in the results of text analyses I’ve conducted on wine back labels using data from Wine.com:

    http://markallenthornton.com/blog/wine-back-label/
    http://markallenthornton.com/blog/price-of-flavor/

    I’d be happy to hear your thoughts!

    Mark

  3. Comment on the viognier label. I’m always offended by misspellings and punctuation errors, even on a wine label. Is it Clarkburg or Clarksburg, and where did they get all those misplaced commas from. Viognier seems to be spelled correctly even if the spelling checker disagrees.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: