Onomastic findings, especially in toponomy, are of fundamental importance for the assessment of a geographical continuity regarding settlements that outlasts language changes (cf. Krefeld 2020c, Krefeld 2021i and especially regarding the Alps Steub 1867 and von Greyerz 1933). Here, two semantic types can be distinguished: On the one hand, ancient forms from substratum languages were being adopted ('-these, often unpleasant, tattlers from past times', as Steub 1867, 142, remarks smugly). During their retrieval, the regularity of historical phonetics has to be respected while, at the same time, various exceptions – often caused by folk-etymological renovations – can occur. On the other hand, there are 'speaking' names which were used to explicitly denominate speakers of substratum languages, like, e.g., the Alemannic and Bavarian names for 'Romansh' with the base wal(l)-, walch-, welsch- (c.f. the classical work by Ernst Schwarz 1970 as well as this map). The procedure can be approached from two opposite perspectives. Either you search for a current or more recent equivalent based on ancient documents (like, i.a., in itineraries, on milestones) or you reverse the direction by identifying etyma for current forms by means of ancient documents or linguistic reconstructions.
The first-mentioned perspective is undeniably simpler because of the relatively small amount of input data. It offers a basic set of Roman names that are mostly easy to locate and that can be worked through systematically. The reverse perspective is far more demanding and also considerably more informative. What is missing to large extent here is fruitful, large-scale research. Many areas have not been studied onomastically since neither the recent dialectal variants nor early records are accessible. At least, a promising first step has been made by the Swiss toponymy portal.
Apart from the names of settlements, the documentation of area names (like mountains and mountain pastures) is especially important because of the conclusions that can be drawn from them about the continuous use of the terrain. Here is a preliminary example: In his crucial study on early medieval settlement in southern Upper Bavaria and the connecting areas of Tyrol and Salzburg, Franz Weindauer 2014 writes about the results of a comprehensive evaluation of the archaeological discoveries: 'Comparing the results of typonomy and of the study of patrocinia has yet again shown that the regions populated in the Late Roman period deliver credible clues regarding Roman life in the Early Middle Ages. These primarily include the Lake Ammer region, the 'Werdenfelser Land', the Chiemsee area and the Rupertiwinkel, but also, for instance, the regions around the Mangfallknie, Rosenheim and Lake Starnberg' (Weindauer 2014, 249). The arrangement of the archaeological sites in the Alpine part of his research area clearly follows the pass trails, like, e.g., the Brenner route which (viewed from the south) divides into a western branch in Innsbruck and an eastern branch (down the river Inn). Even a cursory topomastical glance at the vast Karwendel mountain range, which lies between these two routes and does not have a single larger settlement to this day, brings to light some evident romanisms or Romansh words influenced by pre-Roman at distances of just a few kilometres:

* the Fereinalm < Lat. veranum, cf. Spa. verano 'summer' as well as the Vereina tunnel in Lower Engadine;
* the Krapfenkarspitze based on pre-Roman *krapp- 'boulder', cf. Roh. (sursilvan) crap with very numerous records in Grison toponomy (cf. Schorta 1964 , 111–114);
* the Pleisenspitze with the pre-Roman base *blese 'steep grass slope', cf. Roh. blaisch, blais, bleis, bleisa (cf. DRG 2, 373 as well as the equally numerous toponymical records in Schorta 1964 , 44–46);
* the Hochgleirsch Lat. glarea 'gravel' + iciu with a shift of the stress to the first syllable; cf. the Swiss equivalents, like the Roh. place name glaretsch in Disentis (cf. Schorta 1964, 164) and the eastern Swiss Alemannic place name Glaretsch in Pfäfers (cf. ortsnamen.ch);
* the Larchetalm < Lat. laricetum 'larch forest', from larix + etum, cf. the numerous Roh. records of the type laret, on the Lat. version larictum in Schorta 1964, 185;
* the mountain name Juifen < Lat. iugum 'saddle', with quite a number of widespread parallels, cf. the place names Juferte in the Simmental south of Bern, multiple Juf and giuf, giuv in Grisons (all of which can be found in the Swiss typonomy portal) as well as the Jaufenpass in Southern Tyrol; slightly to the east of the Juifen lies the mountain Guffert, whose name lines up with the above-mentioned Simmental variant.
A thorough analysis would yield significantly more data. Note the clear shift from Latin to Roman phonetics (-g- > -v-, -f) in the last example. All of these adoptions require contact to a Romansh-speaking population that used the mountains for its subsistence as well as acculturation to the local way of life. This was already expressed by Gamillscheg 1935, 306: 'The Alemannic and Bavarian farmer who laboriously produced fruits from the newly planted soil forms the connection with the Romansh people, working under equally harsh conditions. Nothing shows more clearly that no national struggles have taken place here than the fact that the old Romansh place names and the more recent German ones continue to coexist just as peacefully as Ladins and German Tyroleans. The arrival of the Alemanni and Bavarians in the Rhaetian-Noric Alpine region did not result in the destruction of culture.'