Abstract
The words of a language are randomly replaced in time by new ones, but longsince it was observed that words corresponding to some items (meanings) areless frequently replaced then others. Usually, the rate of replacement for agiven item is not directly observable, but it is inferred by the estimatedstability which, on the contrary, is observable. This idea goes back a long wayin the lexicostatistical literature, nevertheless nothing ensures that it givesthe correct answer. The family of Romance languages allows for a direct test ofthe estimated stabilities against the replacement rates since the protolanguage(Latin) is known and the replacement rates can be explicitly computed. Theoutput of the test is threefold: first, we prove that the standard approachwhich tries to infer the replacement rates trough the estimated stabilities issound; second, we are able to rewrite the fundamental formula ofGlottochronology for a non universal replacement rate (a rate which depends onthe item), third, we give indisputable evidence that the stability ranking isfar to be the same for different families of languages. This last result isalso supported by comparison with the Malagasy family of dialects. As a sideresult we also provide some evidence that Vulgar Latin and not Late ClassicalLatin is at the root of modern Romance languages.