We present the first attempt to construct a long-run historical measure of subjective wellbeing using language corpora from millions of digitized books for the USA, UK, Germany, France, Italy and Spain. While existing measures go back at most to the 1970s, our measure goes back at least 200...