This paper presents ongoing work with automatic transcription of handwritten, phonetically precise dialect texts from the south-west of Sweden (collected in the 1890s). Using a SAMPA-based transcription key (where SAMPA stands for Speech Assessment Methods Alphabet), I have enabled the training of an HTR engine (where HTR stands for Handwritten text recognition), by feeding it manual transcriptions, to automatically transcribe two separate (but similar) phonetic hands. The phonetically detailed output reveals structural properties of the dialect that are hard (at best) or impossible (at worst) to retrieve from other sources. In this paper, I show how my research on enclitic pronouns in North Germanic has benefitted from the possibility to search for prosodic dependencies that the digital versions of the dialect texts provide.