Thursday, December 11, 2014

What language tells us about the roots of the stone age diet

As a restaurant offering ‘paleo’ food opens in London, can linguistics shed any light on what our ancestors ate?

Next month, the UK’s first paleolithic restaurant will open in London. The paleo, stone-age or caveman diet has been around in various forms for decades. But it harks back to a much earlier time, when our eating habits were supposedly in line with our evolution, before agriculture came along and made us civilised but unhealthy.

One of the best TEDx talks I’ve watched is an efficient debunking of the stone-age diet by Christina Warriner. In short: the meat-heavy regimen is not at all how cavemen ate, and many so-called stone-age foods – like root vegetables – are in fact the product of intensive cultivation (wild carrots, for example, are generally tiny and full of toxins).

Things we know about our ancestors generally come from two sources: their own accounts, and the things they have left behind. In other words, history, and archaeology. When there’s no written record of a culture – and the earliest examples of writing are around 5,000 years old – we have to rely on artefacts and biological remains. Authors of stone-age diet plans have relied on out of date and, well, half-digested archaeological evidence of what life was like thousands of years ago. Accuracy isn’t their priority – selling books is.

Strangely enough, though, there’s another line of inquiry available. Since the mid-20th century, scholars have argued that language can provide clues about stone-age lifestyles, even in the complete absence of writing. How could that be possible?

It’s all thanks to a painstaking technique called the comparative method. This is as archaeological as linguistics gets: it involves delicately scraping away the surface layers of spoken language to find patterns of relatedness beneath. For example: the word for “father” is pere in French, but far in Danish. Now, we know these two languages have common roots, and in that context two forms of a word as basic as father will almost certainly related – or “cognate”. We also know there are many other words that start with a “p” in French but “f” in Danish, like pied andfod, poisson and fisk. At some point in the past, then, a change must have crept through Danish and French to make them less alike in this small way. Did a “p” become an “f”, or vice versa? Because we also know that the former is a very frequent kind of sound change, we can deduce that the common ancestor of both French and Danish used a “p” at the beginning of these words.

With a highly technical understanding of sound change, and collections of thousands upon thousands of cognates, whole words, and then large vocabularies of unrecorded languages have been revealed. Much of the early history of academic linguistics was spent reconstructing European “proto-languages”, our best guesses at earlier forms of familiar linguistic groups. There was proto-Germanic, proto-Romance, proto-Slavic and, eventually, a huge lexicon of proto-Indo-European (PIE) a hypothetical language that represents the common ancestor of everything from Welsh to Romanian, Greek to Sanskrit.

This, of course, takes us back quite a long time. Just how long is a matter of fierce debate. In all probability it gets us just a bit beyond the earliest writing, about 6000 years ago. Not quite palaeolithic then, but late neolithic – still stone age, and before, or at the very dawn of agriculture. What did society look like then? How did people live, eat and hunt? Well, let’s look at some of the vocabulary we’ve been able to reconstruct (thanks to the database over at University of Texas, Austin, for these examples)

  • bol- (root) 
  • bhabha (bean)
  • kerem- (onion)
  • abel- (apple)
  • rktho-s (bear)
  • ghan-s- (goose)
  • anut- (duck)
  • kuon- (dog)
  • ulp- (fox)
  • bhren-to-s (deer)
  • ghaiso- (arrow)
  • aik- (spear)
  • deru- (to work, toil)
  • gelebh- (to flay, skin)
  • bhes- (to rub)
  • ten- (to stretch)
  • plek- (to plait)
  • kista (basket)
  • médhu (mead)

My list is highly selective. But this does sort of sound like a subsistence culture in northern Europe. The PIE diet might include some deer stew, with onions and beans. But significantly, there’s no olive oil, fig trees or citrus fruits.

Which leads me to my favourite map in all of linguistics. It’s based on the idea that, because we have been able to reconstruct words for apple, salmon, oak, beaver, squirrel, hedgehog but not grapes or chestnuts, we can work out where the Indo-European “homeland” might have been.

Sadly, there is a big problem with this approach. The PIE lexicon is far from complete. What is more, names can be applied differently by different cultures. Words related to salmon have also been used to indicate trout, a fish with a different range. And meanings are slippery. What if we took the existence of a word “mead” to mean that PIE speakers knew about fermentation and alcohol? Our word beer, after all, comes from the medieval Latin biber – a drink of any kind. Not only that, but at 6,000 years’ distance, the linguistic ground is far from firm. The datasets get smaller and smaller the further back you go, making it harder to say that similarities between words aren’t down merely to chance. Linguistic palaeontology promises more than it delivers.

Still, it’s nice if language adds some colour to the blurry picture we have of our distant forebears. And you could probably do worse than follow a PIE-diet: no short-crust pastry, but oak-smoked hedgehog in abundance.
Shariatmadari, David. 2014. “What language tells us about the roots of the stone age diet”. The Guardian. Posted: October 22, 2014. Available online:

No comments: