¹ Lacking a scientific definition of ‘standard German’, I have made an arbitrary one for the purposes of this paper. I have based my investigations on the speech of a single individual, a member of the upper middle class and a native of northern Germany. It is the speech of my wife, Jenni Karding Moulton, to whom thanks are due for her infinite patience as an informant. She was born in Flensburg, Schleswig-Holstein, but moved at an early age to Berlin, where she lived until 1938. She is one of those rather numerous middle- or upper-middle-class northern Germans who are almost totally ignorant of any Low German local dialect, having spoken only the local variety of High German from earliest childhood. Her knowledge of Low German is limited to a few jokes, songs, and poems. As an informant for ‘standard German’, such a speaker from the Low German dialect area is perhaps preferable to one from the High German dialect area. Speakers from the latter area usually grow up speaking the local dialect; their ‘standard German’ is usually just a more or less extensive modification of the local dialect.

A few peculiarities of the informant's dialect should be mentioned. She is a /╹funt/ Pfund ‘pound’ speaker; I have standardized this to /╹pfunt/, etc. Where standard German demands long vowel plus /ren/, her speech shows either /rn/ or /ern/, in free variation: thus führen ‘lead’ is either /╹fü:rn/ or/╹fü:ern/; I have standardized this to /'fü:ren/, etc. Unvoiced /g/ is of course nearly always /x/: /╹tra:gen/ ‘carry’, but /'tre:xt/ ‘carries’, /╹tru:x/ ‘carried’, etc. Long vowels are often shortened before unvoiced /b d g/: /╹gra:ben/ ‘dig’, but /das╹grap/ ‘the grave’; /╹ra:deln/ ‘(to) bicycle’, but /das╹rat/ ‘bicycle’; /im╹tsu:ge/ ‘on the train’, but /de:r╹tsux/ ‘the train’.

W. Freeman Twaddell was kind enough to read an earlier draft of this article. I am indebted to him for a number of suggestions and corrections.

² [a] is a higher-low central unrounded vowel. Note that the [a] of [štAt] ‘city’ is both shorter and higher than the [a:] of [šta:t] ‘state’.

³ My informant has an eighth long vowel, roughly [ϵ:], which is used as the name of the letter ä, and as a mock-formal pronunciation for words in which [e:] is spelled ä. Thus, spät ‘late’ is normally [špe:t]; but a mock-formal pronunciation is [špϵ:t]. This may, if we wish, be analyzed as a marginal phoneme /ä:/.

⁴ Plus [uj] in the one word [pfuj] ‘phooey’.

⁵ Non-syllabic [r], syllabic [ṛ], is a lower-mid unrounded vowel, halfway between central and back. (Limitations of the type font have led to the choice of this symbol.)

⁶ Also, if we consider utterances of more than one syllable, in [╹li:njə] ‘line’, [t'wA╹lϵtə] ‘toilet’, [╹janwa:r] (beside [╹janu:a:r]) ‘January’, [zi:twa╹tsjo:n] (beside [zi:tu:a╹tsjo:n]) ‘situation’, etc.

⁷ Aspiration is here used as a cover-term to include voiceless vowels, voiceless [l], voiceless [n], voiceless [γ] (= [x]), and voiceless [v] (= [f]). Strictly speaking, we find the following combinations (where [k] stands for [p], [t], or [k], [a] stands for any vowel, and voiceless sounds are indicated by superior letters): [k'a] = [k^aa]; [k'la] = [k╹la]; [k'na] = [kⁿna); [k'γa] = [k^γγa]; [k'va] = [k^vva].

⁸ Pitch is also meaningfully distinctive in German, but I have not fully investigated it.

⁹ It will be noted that the three degrees of stress which I have called ‘normal’ are used to distinguish morphemes: /ˀain/ ‘a’, /╻ˀain/ ‘one’, /╹ˀain-/ ‘in’; whereas the abnormal (overloud) degree of stress is used to distinguish whole utterances.

¹⁰ The number of absolute degrees of stress is theoretically infinite. An even weaker degree than what I have called ‘normal minimum’ is easily distinguished: [ˀn̥╹flus]. But it is not meaningfully distinctive, since the meaning is still ‘a river’.

¹¹ The material in §§2.21–2 might be re-stated in terms of syllable division; i.e. instead of speaking of ‘onset of stress’, we might speak of ‘syllable release’. In /╻kau╹flaks/ the second syllable is released by the /f/; this distinguishes it from /╻kauf╹laks/, in which the second syllable is released by the /l/. This would mean that the place of release of a strongly stressed syllable is phonemic; and we could re-define the loud stress phonemes as follows: /╹/ is not just ‘loud stress’, but ‘place of release of a loud stressed syllable’; and similarly with /╹╹/ and /╻/.

The reason I do not choose to speak of syllable division is that I am often unable to determine the place of release of an unstressed syllable. In the case of such an utterance as /ˀinber╹li:n/ ‘in Berlin’, I am reasonably certain that the second syllable is released by the /b/; but I suspect that this is only because I know that the cluster /nb/ occurs neither at the beginning nor at the end of an utterance. But in the utterances /ˀinsla╹vo:nien/ and /ˀinsla╹tainiše/, I am unable to determine whether the second syllable is released by the /s/ or by the /l/. (The cluster /ns/ occurs at the end of an utterance, as in /╹gans/ ‘goose’; and the cluster /sl/ occurs at the beginning of an utterance, as in /sla╹vo:nien/ ‘Slavonia’.) It may well be that there is a physiological difference in the articulation of the segment /ˀinsla/ in these two utterances; I wish I had the training in the use of laboratory instruments to find this out. But there is no audible difference; and any physiological difference can therefore not be phonemic.

¹² Also excluded are such isolated sound sequences (also foreign borrowings) as [a:j] in [de:╹t'a:j] ‘detail’, [wi] in [╹wiski] ‘whiskey’, etc.

¹³ Cf. Bernard Bloch, Studies in Colloquial Japanese I: Inflection, JAOS 66.98 footnote 4 (1946). I am indebted to Bloch for the suggestion that /+/ be analyzed as a segmental phoneme with allophones consisting of pause and zero.

[This is a correct interpretation of the statement in JAOS 66.98; but there is a better way to explain the analysis of [šu:] as /syuu/ and of [ča] as /tya/: the sound [š] is simultaneously an allophone of /s/ and an allophone of /y/; the sound [č] is simultaneously an allophone of /t/ and an allophone of /y/. In general, a sound belongs to as many phonemes as it contains distinctive features—here voiceless sibilance and palatalization in [š], voiceless dental occlusion and palatalization in [č]. (The order /sy/ rather than /ys/, etc., is established on the basis of patterning: thus there is /ky/ but no /yk/, etc.)—BB]

¹⁴ The fact that /+/ occurs almost exclusively at syntactic and morphological boundaries raises the question: Should we accept syntactic and morphological boundaries as part of our phonemic analysis if, by so doing, we can limit the scope of—or even avoid assuming—open juncture?

For a number of reasons I believe that this should not be done. First—and this is a purely methodological reason—I believe that the phonemes of a language should be analyzed without reference to syntax or morphology (as I have tried to do in this paper). Secondly, we could not do so successfully even if we tried, because of the cases (noted above) in which open juncture does not coincide with a syntactic or morphological boundary. Finally, it would seem that the phonetic marking of morphological and syntactic boundaries is more clearly described precisely by the assumption of open juncture.