The distribution of the forms belonging to this base type is not easily assessable. This is because the Standard German Keller can also be traced back to the Latin cellārium, so that the question arises whether the Alemannic and Bavarian attestations in the study area should be seen as variants that came with Standard German, or whether they should be considered relics of the Latin-Romanic substrate. The argument in favour of the substrate is undoubtedly semantics, because in the Germanic Alpine region, as in Romansh, the dominant meaning is 'milk room, room/hut for storing milk and cheese' or 'hut for processing milk'. This meaning is primarily functional and defined by purpose rather than architectural aspects and therefore corresponds much more to the Classical Latin meaning of cellārium, namely 'pantry, larder' than to the meaning 'basement' in Standard German Keller. Italian cellaio also tends to denote the storeroom; the 'basement', on the other hand, is called cantina. The Romanic records thus show an ethnographically evident, slight semantic specialisation. The development of the meaning from 'storage room' to 'cellar' is also very plausible, especially in the case of wine, which is often stored in the cellar. On the other hand, the regression from deu. Keller to 'storage room for milk and cheese', i.e., precisely to the presumed old meaning of the neighbouring Romanic forms, is very unlikely.
However, the phonetics of the Alemannic and Bavarian forms are complicated, as they show no reflex of the Romanic palatalisation of the initial [k-]. This problem arises not only for southern Germany, but for the entire early Latin-Romanic/German borrowing area, as shown by the juxtaposition of the shifted (deu. Zwiebel < lat. *cēpŭlla; REW, 1820 under cēpŭlla) and unshifted forms (deu. Kiste < lat. cĭsta 'basket', deu. Wicke < lat. vĭcia). Note in this context the name of the river deu. Neckar < lat. Nicer (cf. RE, XVII,1 and dKP 4, 88) without any palatalisation. This name was borrowed with a certain probability before 260–280 AD, since the areas of Germania superior on the right bank of the Rhine, including the entire course of the Neckar, were abandoned in this period. Thus arises a terminus post quem for palatalisation in the Northern Alpine Empire or, in more cautious terms, for its general implementation. Given the fundamentally old age of Romanic palatalisation, it is not convincing to argue only with the time of the borrowing. Rather, one should consider that unshifted, conservative and shifted, innovative variants coexisted in Early Romanic over a long period of time. Note that the plosive did by no means survive only in the early romanised, isolated and quite distant Sardinian (cf. the well-known examples such as srd. kentu 'hundred' \ lat. centu[m] etc.), but also seems to have existed in Dalmatian. In this case, the distance to Alpine Romanic is no longer very great (cf. Dalmatian kapula < lat. *cēpŭlla; REW, 1820 under cēpŭlla).