[skip to content]

Department of Linguistics

Mr Shanti Ulfsbjorninn

BA (SOAS), MPhil (Cantab)


Shanti Ulfsbjorninn
Department of Linguistics

Teaching Fellow

Mr Shanti Ulfsbjorninn
Email address:
SOAS, University of London
Thornhaugh Street, Russell Square, London WC1H 0XG
Russell Square: College Buildings
Office No:
Thesis title:
The Role of Empty Nuclei in Stress Systems
Internal Supervisors

PhD Publications

  • 2010. C-Command, Projections and Melody: Microparameters of the Han-Template. Kaye, J. D. and K. Balogne-Berces (eds.). Acta Linguistica Hungarica. 57(4):x-x.
  • 2009. [invited]. The Internal Structure of Compounds: A phase account of aphasia. Iberia. 1(1):143-168. www.siff.us.es/iberia/index.php/ij/article/viewArticle/7
  • 2008. The Formal Dynamism of Categories: Fricatives vs. stops, simplicity vs. primitivity. Dziubalska-Kolaczyk (ed.). Poznan Contemporary Studies in Linguistics. 44(1):121-146. versita.metapress.com/content/65468526279vn752/
Online Articles (Rutgers Optimality Archive, lingBuzz):
  • 2008. MOVE as AGREE and I(nternal)-MERGE in Phonology. (ms.). University of Cambridge. lingBuzz/001012. ling.auf.net/lingbuzz/@FTrstXHNLbIAexgU/JCNAvfyl?4
  • 2007. Floating Consonants in French: the need for skeleton at input. (ms.). SOAS – University of London. ROA/891-0107.


My job is to show how Empty Nuclei interact with the Stress systems of the languages of the world, specifically I aim to prove the following:

  1. Where the phonemes start and stop does not necessarily equate to the edge of a phonological-domain.
  2. We cannot learn much (if anything) about the placement of fixed stress based on the current typology because the generalisations were made considering vowels only. Vowels are only a subset of the total number of nuclei in the phonological domain. Empty nuclei may not be able to hold stress, but they are able to count as dependents, hence stress placement analyses based just on vowels lead to incomplete (or wrong) generalisations making the WALS maps on stress almost entirely useless.
  3. In Quantity-sensitive systems the coda cannot contribute to weight, because it does not exist. Our addition to this point, which has already been made for Latin by Scheer and Szigetvari (2005), can this be successfully done with other languages? What about more complicated weight-typologies, such as Hindi. Does positing empty nuclei here simplify our analyses or complicate them?
  4. Codas are not moraic, so why not go the whole hog and claim that onsets cannot be moraic either? That is what I will do by analysing Piraha, Pattanai Malay, Aranda and the language I can't remember the name of now that shows onset-weight word-minimality effects. In all of these languages you would want to posit empty nuclei for totally independent reasons (phonotactics, diachrony, alternations and minimal word-effects), the positing of these empty nuclei accounts for their stress patterns without any recourse to onsets (yes, even in Piranha!). So if codas can't bear morae and onsets can't bear morae, there are no morae there are just nuclei.
  5. After all of that, one ends up with a set of stress patterns that are attested and common and another set of stress patterns that I implicationally claim cannot exist. Absence of evidence is not evidence of absence, however, so I will run an experiment. I come from a 'realistic' view of phonology which means that (a la Smith and Tsimpli 1995) if something is banned by UG then it shouldn't just be rare we should actually not be able to learn it naturally, as an L1 learner learns.

The Languages I am most focusing on this term are: Pulaar, Hindi, Piraha, Chuukese, Leti, Pattani Malay and Italian.

Related work by others...
  • Charette, M. (2008) Fixed Stress (Penultimate) Analysis of Turkish
  • Harris and Gussmann (2002) Analysis of Spanish as a fixed stress language (Penultimate)
  • Scheer, T. and P. Szigetvari (2005) deconstruction of Latin codas into [CVCv] and counting empty nuclei