Read on Twitter
Skirgård et al (2017) published an interesting study on people's perception & recognition of languages on the basis of 'listening' audio i.e. why are some languages confused for others? Here in this Neighbour Net Analysis map the shortest route between 2 languages indicates...
how often they are confused from one another. The further you have to move, the more different two languages sound to the listeners. The data include 15 million guesses (originating mainly from Western countries) from 400 audio recordings of 78 languages.
As you can see in the map, Hindi was often confused with #Gujarati and #Malayalam. #Kannada was often confused with #Tamil and vice versa. Surprisingly #Urdu was quite remarkably distinguished far from Hindi (probably due to highly Persianized text of the "news" audio recording).
#Nepali was found closer to #Bengali, i.e. often confused. #Punjabi was found closer to Malayalam than #Hindi (very surprising !!). One of the key findings of this research was - languages are more likely to be confused if they are similar in their phoneme inventory and Lexicon.
To summarize, despite obvious mismatch btw actual linguistic affinities among Indian languages & how they are juxtaposed by listeners in this study, this report points out how several of English reading western individuals have very different perceptions about Indian languages.
