The classical task: From singular to plural form in Dutch, Danish, Austrian German, and Hebrew

Steven Gillis, Agnita Souman, Sim Dhollander, Inge Molemans, Laila Kjærbæk, Katja Rehfeldt, Hans Basbøll, Claus Lambertsen, Sabine Laaha, Johannes Bertl, Wolfgang U. Dressler, Naama Lavie, Ronit Levie, Dorit Ravid

Research output: Contribution to conference without publisher/journalConference abstract for conferenceResearch


Aims and background: The aim of the ‘Classical Experiment' was to test children's knowledge of plural formation using a set of existing words, unlike a WUG test. Hence we will get a diagnostic view of their suffix selection as well as their knowledge of stem changes. In addition, the results of the experiment will give a picture of how children's knowledge develops over time and how well the dimensions ‘suffix predictability' and ‘stem transparency' are able to account for children's behavior.



Subjects: For each of the four languages, 20 subjects were tested in 7 age groups, ranging from 3-year-olds to 9-year-olds. The experiment was also administered with a control group of adults.

Procedure: The procedure consists of a classical set-up in which the subjects are shown pictures of objects. On presenting each object, the test leader says: "Here is a X (name of the object)". The next picture shows two (or more) instances of X and the test leader says: "Here are two/a lot of ...." and the child is expected to provide the plural form of X.

Stimuli: The target words used in the experiment were easily picturable and judged to be known even by the youngest participants. Moreover, all the test words were found in child addressed speech, see the longitudinal data.


Predictions: Given that the model underlying the design of the experiment was identical for each language, we predicted that the outcome would be identical for each language. More specifically, the following predictions were articulated:

Prediction 1: a global analysis of the plural forms provided by the subjects is expected to show an increase of the correct responses as children grow older.

Prediction 2: As to suffix selection, we expect that the plural of nouns selecting a fully predictable suffix will be more readily mastered than the plural of nouns with a partially predictable suffix, and plurals that are exceptional as to their suffix selection will be most difficult, and, hence mastered later.

Prediction 3: As to stem change, we expect that plurals with no stem change involved will be ‘easier' in each age group, and hence also earlier acquired than plurals with a slight stem change, which in turn will be ‘easier' than plurals involving a substantial change.

Taken together, these predictions amount to the following: given the matrix described in the "Methodology", stimuli in the upper left corner will, in general, be ‘easier' (more correct scores) and will be earlier acquired. Stimuli in the right cell at the bottom will be the hardest in each age group, and will be acquired later than the stimuli in any other cell.


Results: On the whole the results are in agreement with the predictions, in that for each language the topology of the matrix accounts for (1) the relative amount of correct plurals in each age group; and (2) the relative order of acquisition when the data are analyzed longitudinally.

Original languageEnglish
Publication date2008
Publication statusPublished - 2008
EventThe XIth Congress of the International Association for the Study of Child Language (symposiet: "Why are noun plurals hard to acquire? A multi-task approach") - Edinburgh, United Kingdom
Duration: 28. Jul 20081. Aug 2008
Conference number: XIth


ConferenceThe XIth Congress of the International Association for the Study of Child Language (symposiet: "Why are noun plurals hard to acquire? A multi-task approach")
Country/TerritoryUnited Kingdom


  • Language acquisition
  • Noun plural

Cite this