Stød in Danish proper names – in standard Danish pronunciation

Nina Grønnum (henceforth NG) is a leading Danish phonetician (born 1945), pupil of Eli Fischer-Jørgensen (1911–2010). She has written a small, but very important book on “Stød in Danish proper names”. The great Danish linguist Louis Hjelmslev (1899–1965) – father of the structuralist linguistic theory called Glossematics – said (1951: 12 [1973: 248]):^[1] “The structural analysis of any individual language will present certain special problems around which the whole analysis must be concentrated […] In Danish, the main problem is the stød and the related matter of latent consonants.” Stød is a contrastive syllable-rhyme prosody historically related to the Swedish and Norwegian tonal word accents.^[2] Stød is prototypically non-modal voicing with aperiodic vocal folds vibrations and irregular variation in amplitude, i.e., laryngealization (p. 11).^[3] Acoustically, there is higher Fo at the onset of syllables with stød (p. 11).^[4]

NG uses Basbøll’s (2005, 2008, 2014) Non-Stød Model (henceforth NSM) which departs from the generalization that heavy, i.e. bimoraic,^[5] syllables have stød by default; this has turned the general approach (giving rules for which syllables have stød) upside down and made it possible to significantly improve the traditional account of stød.^[6]

The book is organized in six chapters of which the first three are introductory: chapter 1 (p. 9–10) argues why the investigation is needed, chapter 2 (p. 11–18) presents Danish stød, and chapter 3 (p. 19–25) is about the investigation, including the data bases. Chapters 4 (on place names) and 5 (on personal names) present the results, and the short chapter 6 is the conclusion. This structure is logical, and the reader is well guided through the investigation.

The cover of the book (designed by NG herself) illustrates the problem: it shows traffic signs of two places (villages): Kærup and Kærum where the penultimate stressed syllable in both names has a long vowel [ε:]; but in Kærup this vowel has stød, not in Kærum, why is that? The book gives detailed and convincing answers to this and many other such questions. NG sets the stage of a person speaking standard Danish – typically a Copenhagener of at least middle-age – who disregards his/her knowledge of specific local variants of names – seeing written place names they (with few exceptions) do not already know, on their tours around in the Danish countryside. The investigation also includes personal names, both boys’ and girls’ names, and surnames.

Methodologically the book deserves high praise for its explicitness and well-argued procedure through all steps of the approach, right from the establishment of the inventory of proper names and their pronunciation. The largest chapter (4) of the book concerns place names (p. 26–51). In November 2020, there were 28,487 authorized place names in the database used (p. 26). NG ends up with about 4,000 place names to be analyzed, by different procedures specified (p. 26–27): she excludes orthographic doublets, and place names containing personal names (that are treated in chapter 5), and also names identical to common words (i.e., with entries in the dictionary Den Danske Ordbog); NG also uses two phonological criteria in the process, viz. excluding names without stød-basis and names with stress on the non-initial part of the name.

NG first notices (p. 28) that the about 4,000 place names are generally not inflected nor derived; she then observes that “disyllabic place names with one of the non-full vowels in Danish phonology in the post-tonic syllable are not complex, they are phonologically^[7] simple structures”. There are about 1,000 simple place names in the corpus (e.g., Ry and Hjallese). The remaining (non-simple) place names are about 1,400 compound-like (e.g., Borup and Hinkbøl), and about 1,500 compounds (e.g., Findal ‘Finvalley’ and Onsevig ‘Onsecove’). NG (p. 28) says that the “last part in a compound-like place name has a phonological structure that could have made it a common word, except that it is without contemporary meaning”, and she gives, e.g., examples like -bøl^[8] in Avnbøl and Ulkebøl (and other generic^[9] parts as -tved and -trup as well). There are about 1,400 compound-like place names. The third main type is Compound place names where the generic part is a common word, as, e.g., -mose ‘bog’ in Bølkemose; there are about 1,500 compound place names in the corpus.

I cannot discuss further details of chapter 4 here (see NG’s conclusion referred to at the end of this book notice), but I refer to four informative tables, the first three departing from the generic (second) part of compound-like place names: Table 1 (p. 36) shows 24 generic parts and their origin and number of occurrences, Table 2 (p. 38) shows stød in 921 disyllabic names, Table 3 (p. 44) shows stød in 479 trisyllabic names; Table 4 (p. 47–48) shows common words as generic parts in 1,499 compound place names.

The other large chapter on names is chapter 5 (p. 52–66) on personal names, divided into boys’ names (p. 52–56), girls’ names (p. 56–58), and surnames (p. 58–66). Whereas most place names would not be known by a Copenhagen standard speaker travelling around in Denmark, many personal names would be known by him/her, and, contrary to the place names, many belong to the non-native(like) part of the vocabulary. In agreement with the NSM, e.g., German names stød-wise act as native names, whereas French and English names do not, i.e., they are stød-less, but they get stød when pluralized with <-er>. Thus, the surname Deleuran danifies its French pronunciation (NG p. 58 transcribes it with voiceless /d/ and a final velar nasal); if this name is pluralized, it adds <-er>, and the final velar nasal will get stød.

This leads us to the (variety of) pronunciation of all the names in the book. I am also a speaker of the Copenhagen standard speech that is used throughout the book, i.e., my speech is rather like NG’s speech.^[10] But in many cases, I would have a different pronunciation of the examples, and other Copenhagen standard speakers would have a pronunciation different from both NG’s and mine. But such differences do not change anything substantial in the results.^[11] This is also because NG has defined the phonological context for, e.g., the examples of place names sharply and economically: what counts is only stød versus non-stød in the stressed syllable (i.e., in the specific (initial) part).

Since the book has used the Non-Stød Model as its theoretical foundation, among its main results is the testing of the NSM as a descriptive and explanatory tool to account for stød in proper names in standard Danish. In the abstract (p. 2), NG says “The results do not challenge the validity of the model; but they suggest an amendment and a differentiation of the perception of boundaries between elements in compound and compound-like names.” In the conclusion (p. 67–68), NG mentions four results of her investigation regarding the NSM:

1. “a non-full vowel in a post-tonic syllable in the lexeme triggers non-stød irrespective of the number of post-tonic syllables.” NG here gives Hedense and Gåbense (both with two schwas) as examples. This is an important amendment of the first word-structure principle of the NSM (out of two), and I accept that without reservations.

2. “Segmentally heavy post-tonic syllables are more prone to yield stød than segmentally lighter ones.” NG here gives the personal names Albert and Wilkens, both with stød on the stressed (first) syllable, as against Hjalte and Brian without stød.

3. “Any monosyllabic specific part with stød-basis has stød in place names ending in rup, drup, strup and trup”. NG here gives, e.g., the examples Sindrup and Møldrup, both with stød whereas the place name Sindal and the compound noun mølpose ‘moth bag’ do not have stød (on their stressed syllable). This induces “a cognitive factor in the assignment of stød, a perceived difference in cohesion between the two parts”, and NG explains that the boundary in the place name Møldrup is weaker than the boundary in the compound noun mølpose which causes the final sonorant in møl to be word final.^[12] This is a very significant result, and I completely agree with NG’s analysis. NG concludes: “I ascribe this connectivity of rup, drup, strup and trup – that distinguishes them from all other generic parts – to their massive preponderance in Danish place names.”

4. “The tendency in common words for those ending in /ən/ to have stød less frequently than those ending in /əl/ and /ər/ is apparent also in names.”

NG adds that “the results can and should be subject to empirical testing”, and she suggests several concrete experiments, the most interesting being one investigating the psychological reality of boundary strength related to point (iii) above.

The book reviewed here is new and innovative but presents a challenging read. It presupposes phonetic and phonological knowledge and an interest in learning. However, readers who are interested in prosody and like to deal with methods and analysis techniques will derive significant benefit from this book. No knowledge of Danish or other North Germanic languages, or of specific laryngeal mechanisms, is presupposed. Thus, any serious reader who is interested in the subject area covered by Phonetica will be rewarded for studying the book.

To conclude: Nina Grønnum has given us a very original and systematic book on stød in Danish proper names. Our knowledge of the relation between phonological word structure and the occurrence of stød has significantly increased by Grønnum’s scrutinizing data that has never been as systematically studied before. Grønnum’s amendment of the first word-structure principle of the NSM (cf. Hedense and Gåbense, see (i) in the conclusion) is considerably strengthened by – but not crucially dependent on – the investigation of proper names. But her results about the effect of the second parts rup, drup, strup, trup of place names on stød in the first part (e.g., Møldrup, see (iii) in the conclusion) could only be reached by a systematic investigation of proper (place) names, as done by Grønnum in this book. And this result is theoretically truly significant.