¹ G: German; Gic: Germanic; H: high; M: middle; O: old; W: west.

Author's name and date replace the full citation of a title in the following cases: the study is listed in Mentz, Bibliographie der deutschen Mundartenforschung, Leipzig 1892 (= Sammlung kurzer Grammatiken deutscher Mundarten 2) or in Martin, Bibliographie zur deutschen Mundartforschung und -dichtung in den Jahren 1921-1926, Bonn 1929 (= Teuthonista, Beiheft 2); the study appeared in ZfdMa, Teuthonista, or ZfdMundartforschung before 1938; the study was reviewed in one of these journals or listed in Jahresbericht fgPh in the year of publication or in the next following year. As the studies are mostly short, and the treatment of [k-] easily found, I have not cluttered the text with page references.

Forms in brackets are normalized from the transcriptions of the various authors, or else are normalized OHG; starred forms in brackets are typical G, with unshifted consonants and weakening of unstressed vowels, or else typical pre-G.

² Of the long fortis spirants which replaced postvocalic [p, t, k], only one was a new phoneme: [Z] from [t], as in [eZZan] ‘eat’; the other two, [ff] from [p] and [hh] from [k], coincided, so far as one can see, with sounds already in the language: [offan] ‘open’ like [heffen] ‘heave’ < *[haffjan], and [mahho:n] ‘make’ like [hlahhen] ‘laugh’ < *[hlahhjan]. New phonemes arose through (later) shortening of these long spirants, [sla:fan] ‘sleep’ contrasting with [gira:vo] ‘reeve’ and [bru:xan] 'brook’ with [li-han] 'leihen’. In the later development, the spirant of [ts] seems to have resembled the new [Z] rather than the old [s].

³ Needless to say, we follow the scheme mechanically. However, I have included references to Vienna among Types 1 to 6, even though it is an island of euch in the area of enk. Wrede's scheme raises a few difficulties of this kind. Thus, among the ich, Apfel, Kind dialects (Types 10 to 23), it provides for the combination of Is, Hus (Baden type) with mähet (Swabian type) for a southern region, Type 14; but this combination occurs also farther toward the north. For the ik dialects the scheme is self-contradictory. If one knows the geographic provenience of a dialect to begin with, this does no harm; for an ik dialect spoken, say, in Iowa, Wrede's scheme would not work. Doubtless, the discussion of the scheme promised for a forthcoming number of Deutsche Dialektgeographie will clear up such matters. Whoever has tried to classify dialects will appreciate Wrede's skill. The divisions make solid blocks and are based on common forms and on differences recognizable even in the poorest transcription.

Even so, we shall sometimes go wrong in placing a dialect. Strange as it may seem, extensive essays about a dialect sometimes fail to give us such forms as us or you or the imperative be. Powerful and deep-seated inhibitions block the study of language, especially of one's own language. One seeks means of evasion, and these may in time grow conventional. In the nineteenth century, the recording of G dialects was usually replaced by a meagre listing of correspondences with MHG. Today the convention leans rather to geographic discussion of borrowing; here too the information often runs very thin, with sometimes scarcely a speech form to a page. The genuine record of a dialect, such as we have in Roedder's Volkssprache und Wortschatz, New York 1936, is almost or altogether unique. Yet without the protocol, there is little value in scientific pretensions.

References which deal with more than one dialect may appear under several types.