UABDivulga - Barcelona recerca i innovació

All over the world, we find transparent toponyms like Vilanova (vila nova in the past, ‘new town’), which we understand what they want to convey, and opaque ones, which constitute a mystery. But despite the fact that any inhabitant would like to know the etymology of that strange name of their town and, at a scientific level, toponyms could provide us with information about how our ancestors thought and spoke, the etymological method has been the same since the nineteenth century and, according to our research, presents some weak phases of the research process. This is why we have designed and experimented with a different methodology. The approach is cognitive and geographic, and the initial objective was not to find out the language in which the opaque toponyms were created, but what they referred to and what characteristics the language of the humans who created them had.

Based on the principles of corpus linguistics, a corpus of 180 Pyrenean population toponyms (from the Pallars Sobirà region, Catalonia, Spain) was first constructed, along with quotations of them in medieval documents. In total, there were 464 toponyms in the corpus. Next, each toponym was associated with its morphological (all possible segmentations of the place name) and geographic variables (the elements of the landscape where the settlement of the toponym is located). Mathematical algorithms, field work, interviews with local people, and Geographic Information Systems were used for data collection.

With the variables prepared, statistical filters made it possible to relate the 1,179 segments of the opaque toponyms of the corpus to 133 elements of the landscape, and seven segments were obtained that were significantly related to seven geographical variables: UI (present in Bretui, Mencui, etc.), related to towns with ‘large visual basins, between 2,501 ha and 6,000 ha’; CA (in Caregue, Escalarre, etc.), related to towns with ‘steep rocky walls’; EST (in Esterri, Balestui, etc.) with towns with ‘flat meadows near the settlements’, etc. From that moment, qualitative and quantitative analyses, and a large number of early medieval documents with around 1,600 historical quotations of place names, made it possible to reconstruct five prototypical (models from which there may be variations) old words that were used to name landscape sites: *oil, *esca, *esta, *orte, *one. It was observed that nearby towns are part of toponymic clusters: opaque toponyms that share segments because they refer to the same significant element of the landscape. It was also possible to articulate etymological proposals for some toponyms.

But the most surprising thing is that these old words show great antiquity. Because, on the one hand, they refer to basic and versatile concepts: ‘something that is cut’, or ‘does not move’, or ‘is on top’, etc. And, on the other hand, they are part of current words of European languages from different affiliations, so they are cognates (they have the same origin). Therefore, this linguistic archaeology research reveals that in the opaque European toponyms, considered linguistic fossils due to their permanence over time, prehistoric cognates from a common old language, possibly Proto-Indo-European, could endure.