We are family

Back in August last year, I wrote a blog entry about cognates (It's all relative, 25 August 2012), which referred to new research indicating that Indo-European languages (of which English is one) may be as much as 9000 years old. A friend alerted me to recent research, summarised in The Guardian, that had been undertaken by some of the same researchers. Their work, published in Proceedings of the National Academy of the United States of America (PNAS), takes the links between languages and language families much further. By applying a statistical model to examine the frequency of certain cognates in everyday speech across seven language families, the researchers assert that high-frequency cognates signify a superfamily of languages that may have existed across Europe and Asia as far back as 15,000 years ago.

Having never conducted any statistical linguistic research myself, the methods outlined simply indicate an attempt by the researchers to counter the usual criticisms or shortcomings of this kind of work in historical linguistics. Having said that, the researchers' results left me feeling that actually, the most prevalent, significant cognates – with controls to take account of chance sound associations – they discovered were actually rather predictable. The list of 23 significant cognates identified, in order of frequency, reads as follows: thou, I not, that, we, to give, who, this, what, man/male, ye, old, mother, to hear, hand, fire, to pull, black, to flow, bark, ashes, to spit, worm. Displaying these words in a table detailing variables such as their respective frequency, half-life (the expected time in 1000s of years before one word has a 50% chance of being replaced by a new cognate word) and the part of speech they exemplify is interesting.

In my view, this is because when they are presented together, knowing that seven language families have been considered; the words invite us to search for universals. From a semantic perspective, it doesn't surprise me that personal pronouns, interpersonal relationships and elements of the natural world are represented – since these words are, and have always been, by virtue of their function, essential to human interaction and/or survival for millennia. The researchers may, essentially, only be revealing by statistical methods what we have always believed to be true. If the pronoun we ain't broke, why fix it, for example? But our non-scientific gut feeling then leads us to consider to spit and worm as anomalies in the context of the table, when further historical or anthropological research may provide further insight.     

So for me, having never previously been interested in anything beyond cognates between modern English, German and French, this research will prove to be significant if its statistical model can be replicated and expanded in further studies. For if we are beginning to understand the rate of language change on a global scale from the distant past until the present; then the next step will surely be to apply the methodology to hopefully be better able to predict changes in languages and communication far into the future.

It's all relative

In February 2011, students started telling me they thought English was a "really difficult language to learn" and that German was "much easier". Comparing languages based on their perceived degree of difficulty for the learner is clearly a subjective and relative concept. In his fascinating book Through The Language Glass: Why The World Looks Different In Other Languages, on page 108, author Guy Deutscher makes the point that the relative difficulty of learning a given foreign language crucially depends on the mother tongue of the learner concerned. Most people would rightly assume that learning Italian is easy if your mother tongue is French, for example. But German and English also have many similarities. So in a bid to enthuse my disillusioned students, and thinking back to my 14-year-old self discovering similarities between the two languages; I designed a fun lesson on cognates – words used in different languages that derive from the same form.

So without going into the High German consonant shift (also called the second sound shift) or the first sound shift (also called Grimm's Law, after the linguist and editor of fairy tales Jacob Grimm), charts available online show that English and German are full of cognates. The most obvious of these are the ones whose spelling is essentially intact in both languages: hand (German: die Hand), finger (German: der Finger), ball (German: der Ball), salt (das Salz), pepper (der Pfeffer), light (German: das Licht). The list goes on and on but you get the general idea.

Now, my reason for drawing attention to the similarities was to show that in lots of ways, by sharing many of the same words, functions and ideas; many languages are not as different from each other as they seem. So you can imagine my delight at reading this morning that researchers in New Zealand looking into cognates – using analyses more commonly employed in evolutionary biology – have discovered that the origins of English, along with those of other languages in the Indo-European family of languages (comprising around 449 of the world's languages), were first present around 5000 years ago. The use of a family-tree theory about language is not new. The technique was first conceived by German linguist August Schleicher in 1853. Though having applied it in this new way using 207 cognate words, the team at the University of Auckland now believe that this Indo-European family is rooted in the Anatolian region of what is now Turkey, and as a result is actually around 8000 years old.

From a sociolinguistic perspective, we are aware that language is subjected to many different influences such as loan words from other languages, semantic shifts and cultural change. However, this new research suggests that cognates can be seen as 'linguistic DNA', such that cognates describing key concepts in the physical or natural world (such as body parts) are much more resistant to change over time. As a result, they may prove to be more reliable indicators of a language's origin.