Schwa Epenthesis and Schwa Deletion in Dutch

In 1996, Cecile Kuijpers, Wilma van Donselaar and Anne Cutler from the Max Planck Institute for Psycholinguistics in Nijmegen, Netherlands conducted a pyscholinguistic study on schwa epenthesis and schwa deletion in Dutch. Schwa deletion and epenthesis are optional in Dutch. In those cases in which schwa deletion and epenthesis occur we have phonological variation. Thus the forms without schwa deletion and epenthesis can be considered the standard forms and those with deletion and epenthesis the phonological variants.

The words "melk" (milk) and "elf" (eleven) are examples of words which can be pronounced with an epenthetic schwa. The epenthetic schwa only occurs with heterorganic consonant clusters. This makes sense because more articulatory effort is required to produce consonants which share different places of articulation. In "hals" (neck) and "damp" (vapour) no schwa epenthesis occurs. The word "hals" has two alveolars and "damp" has two bilabials. Since the clusters in those words are homorganic, the schwa epenthesis rule cannot apply.

In the word "kapelaan" (chaplain) schwa deletion can apply. English also has examples of optional schwa deletion. For example, many speakers delete the schwa of "interesting" and "different" to create the consonant clusters "tr" and "fr".

The goal of the auditory lexical decision experiment was to determine whether or not the participants would process the phonological variants as quickly and accurately as the standard forms. Since the phonological variants are optional, the prediction was that the standard forms would be processed more quickly and accurately.

The participants listened to standard realizations and phonological variants of real words as well as pseudowords. They needed to decide as quickly as possible whether the word was real or not. If they did not decide within 1.5 seconds, their response was recorded as missing.

The results only partially confirmed the predictions. The experiment revealed that it was easier to process the standard realizations of words than the phonological variants with schwa deletion but not in the case of schwa epenthesis. In the latter case, the phonological variants were processed as quickly and accurately as the standard realizations. The words with schwa epenthesis were easy to process for participants. The reason is probably due to the fact that the variants with schwa epenthesis had the CVC syllable structure in the coda, a more basic structure than the CCV in the onset of variants with schwa deletion.

Not surprisingly, the real words were processed more quickly than the pseudowords. Recognition of the real words certainly aided the participants in their responses.

The psycholinguistic experiment provided evidence that syllable structure is important in lexical decisions. Words with basic syllable structures are processed more quickly and accurately than those with complex ones. Also, standard forms and real words are processed more easily than phonological variants and pseudowords.