How Google Translate is Revolutionising the Study of Emotions

In the past, anthropologists have struggled to compare societies that are guilt-focused versus shame-focused. Now Google Translate is allowing them to compare the cultural role of these emotions on a global scale for the first time

The Physics arXiv Blog
The Physics arXiv Blog

--

The evolution of emotion and the role it plays in societies is one of the great puzzles in the study of human culture. Anthropologists have attempted to get a handle on this by studying the role of closely connected pairs of emotions. In particular, they have focused on guilt and shame, attempting to classify societies according to whether they are more dominated by one or the other.

That’s difficult, time-consuming work that requires extensive understanding of many societies and the way they use language. Consequently, this kind of classification has only ever been done on a relatively small scale.

Today, that changes thanks to the work of Klaus Jaffe and pals at Universidad Simón Bolívar in Caracas, Venezuela. These guys have used Google Translate to search for synonyms of shame and guilt in 64 different languages. By assuming that societies with more words for an emotion place greater emphasis on it, the results reveal for the first time a clear spectrum “guilt societies” and “shame societies” with some particularly interesting outliers.

First, some basic definitions. Anthropologists focus on guilt and shame because although similar, these emotions differ in subtle but important ways.

In English, shame is the feeling of humiliation due to the consciousness of wrong or foolish behaviour. It is thought to derive from an older word meaning to cover and has clear physiological consequences that are visible to others, such as blushing, sinking the head and hiding the face.

By contrast, guilt is the emotion of being responsible for committing an offense. One clear difference from shame is that it is essentially an internal emotion and does not necessarily have a clear external manifestation.

This difference is crucial. “The uncontrollable trigger of shame signals makes it an honest signal trait which serves to regulate compliance with social norms,” say Jaffe and co. And for that reason, shame plays an important cultural role.

But while clear “shame signals” give these societies an evolutionary advantage, there is also an advantage to not demonstrating these signals. “The lack of shame also has evolutionary advantages as it allows cheating and thus benefiting from public goods without paying the costs,” point out Jaffe and co.

So an important question is which emotion ends up dominating. To find out, Jaffe and co assume that any society will have more words for things it considers more important; just as the Arctic Inuit people have a dozen words for snow while the Amazonian Yanomami have none.

So these guys used Google Translate to search for synonyms of the words guilt and shame as well as related words such as embarrassment, fear and pain.

The results are revealing. They say that almost every language they studied has at least one concept related to shame and one to guilt and most have more than one.

However, 14 languages have words that do not properly distinguish shame and embarrassment. These include French, Catalan and Persian. Slovak has one word that describes both shame and pain while Chinese has different words for concepts such as the fear of losing face, the feeling state after losing face and the feeling state for someone else who has lost face and so on.

By contrast, the Amerindian language of the Yanomami is the only language that uses the same word—’kili’—for shame, pain, fear, guilt and embarrassment.

Jaffe and co describe the word ‘kili’ in the following way: “Examples of the context when they feel “kili” are: a tiger appears in the forest; you kill somebody from another community; your daughter is going to die; everybody looks at your underwear; you are caught stealing; you soil your pants while among others; a doctor gives you an injection; you hit your wife and others find out; you are unfaithful to your husband and others find out; you are going to be hit with a machete.”

One interesting discovery is that linguistic families do not share the same bias towards guilt or shame. For example, Latin languages such as Spanish, French, Hungarian Italian and so on, all differ widely on this scale.

Outliers of interest include Hebrew which has four guilt words and 13 shame words, Azerbaijani which has 6 guilt words and 13 shame words, Latin with 9 guilt words and 8 shame words. By contrast, English has only 3 guilt words and one shame word.

An interesting aspect of this work is that mathematical models of the evolution of these emotions suggest that neither is an evolutionary stable strategy. In other words, a guilt society is always vulnerable to the spread of shame and vice versa.

That is perhaps reflected in the diversity of the languages Jaffe and co study, particularly between the ancient language of Latin (guilt-focused) and its modern descendants of Italian (shame-focused) and Spanish (shame-focused).

That implies that these societies are vulnerable to a further shifts along the guilt-shame axis.

Of course, this is just the start of a new way of studying societies based on the way they use emotion in language. It may even be possible to track the evolution of these emotions directly by examining changes in their use over time using the books Google has scanned.

That makes this area where linguistics meets anthropology meets ethnobiology particulalrly interesting. Lots of low hanging fruit here.

Ref: arxiv.org/abs/1401.1100 : On The Biological And Cultural Evolution Of Shame: Using Internet Search Tools To Weight Values In Many Cultures

--

--