A phonetically-based phoneme analysis of the Danish consonant system

ABSTRACT

The traditional phoneme analysis of the Danish consonant inventory links onset and coda consonants on the basis of historical alternations and morphologically conditioned alternations within a small subset of the Danish lexicon. This traditional analysis proposes a system resulting in a large number of neutralizations that cannot be dissolved, and in which allophones of the same phoneme lack shared phonetic content. We argue that the system proposed by the traditional analysis is impossible to learn from the language input, which renders the analysis an implausible description of the Danish consonant system. On the basis of theoretical discussions, we offer an alternative phoneme analysis, which we believe to be learnable from the data available in the language input. Our analysis is based on insights from Natural Phonology and Bidirectional Phonetics and Phonology. We propose a system without undissolvable neutralizations, with shared phonetic content between allophones of the same phoneme, and without the need to rely on alternations that children are unlikely to learn in early childhood.

KEYWORDS:

• Phoneme analysis
• Danish
• consonants
• learnability
• phonetics-phonology interface
• Natural Phonology
• Bidirectional Phonetics and Phonology

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

¹ An exception is that Danish plosives may be voiced inter-vocalically. This has long been assumed to be essentially a categorical phonological process (Fischer-Jørgensen 1954, 1980; Fischer-Jørgensen and Hutters 1981), but has recently been shown to be much less frequent than previously thought (Puggaard-Rode, Horslund, and Jørgensen forthcoming).

² Note that the aspirated alveolar is affricated and is typically transcribed with a superscript s, [t^s].

³ The so-called soft d is a centralized alveolar-velar semivowel (Juul, Pharao, and Thøgersen 2016; Schachtenhaufen forthcoming; Siem 2019).

⁴ IPA lacks a diacritic for alveolarization, which makes Schachtenhaufen forthcoming transcribe the soft d as [ɤ̻̈] with a laminal diacritic. However, according to Schachtenhaufen (mail 25 November, 2021), apical pronunciations are also found, which has led us to not transcribe the alveolarization of the soft d in the absence of a fitting diacritic.

⁵ Note that [smɑʊ̯] was used in derivations like smagfuld and smagløs in conservative High Copenhagen varieties in the early 1990s (Brink 1991).

⁶ [ð̞] is a central alveolar approximant with a prominent velar constriction (Basbøll 2005), which is a semivowel [ɤ̯̈] in modern Danish (Juul, Pharao, and Thøgersen 2016), cp. section 2.

⁷ See Brink (2011) for some dubious cases, however.

⁸ Fischer-Jørgensen 1967, III. That Lier’s participation was peripheral, according to himself, relies on personal information from Frans Gregersen (mail 3 June, 2021).

⁹ Uldall uses the term ‘implication’.

¹⁰ The rule that complex clusters must always be resolved into existing simpler clusters was established by Hjelmslev in his theoretical paper at the London conference where Uldall’s analysis was presented, cp. Hjelmslev (1936) (Hjelmslev 1973, 164).

¹¹ Published in Danish as Hjelmslev (1951), here quoted from the English translation in Hjelmslev (1973).

¹² Similar minimal pairs with a [k/ɰ] contrast seem not to exist; in all cases that we tested, like myg [myk/myː^ʔɰ] ‘mosquito/supple’, hæk/hæg [hɛk/ hɛː^ʔɰ] ‘hedge/bird cherry’, klæk!/klæg [k^hlɛk/k^hlɛː^ʔɰ] ‘hatch!/sticky’, muk!/måg [mɔk/mɔː^ʔɰ] ‘grumble!/brother or son in law’, or nik/neg [nek/neː^ʔɰ] ‘nod/sheaf’, such pairs also have a distinction between long and short vowel. The present-day reduction of [ɰ] into [ɪ̯] or [ʊ̯] is not taken into account in these transcriptions.

¹³ In words like these, [ɪ̯] is often dropped in fast speech and sometimes also in citation forms.

¹⁴ Which evolved on the basis of Stampe (1969).

¹⁵ Plosives may be aspirated in utterance final position, as may all sounds (Grønnum 2005, 310).

¹⁶ Note that Basbøll (2005, 281, footnote 46) assumes suppletion in cases like god ‘good’ and bedre ‘better’.

¹⁷ In a wholly different context, Basbøll (2005, 138ff.) discusses the feature [grave] as encompassing both labial and velar consonants. He concludes that it is not used to distinguish Danish phonemes, although it does serve to explain the distribution of short /a, æ/ (see Basbøll 1972).

¹⁸ CSH was born in 1986 in Skive, RPR was born in 1992 in Varde, and HJ was born in 1953 on Frederiksberg.

Additional information

Funding

This work was supported by the European Union’s Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie under grant agreement No [894936].

Notes on contributors

Camilla Søballe Horslund

Camilla Søballe Horslund is a postdoctoral fellow at the Department of Communication and Culture at Aarhus University. Her research is primarily within phonetics and (laboratory) phonology and the interface between the two with a focus on (second) language acquisition. She has also published on second language acquisition of syntax and vocabulary and on language change.

Rasmus Puggaard-Rode

Rasmus Puggaard-Rode is a graduate student at Leiden University Centre for Linguistics. His research interests mostly lie in phonetics and (laboratory) phonology, how the two inform each other, and how the use of corpora can improve our understanding of both. He has also published on dialectology, interactional linguistics, forensic linguistics, and second language acquisition.

Henrik Jørgensen

Henrik Jørgensen is a senior lecturer of Scandinavian Linguistics at the Department of Communication and Culture at Aarhus University. He has published on phonetics, morphology, syntax and semantics, and he is currently the leader of a project on Danish structuralism in the 20th century (Infrastrukturalisme).