abstractness classification

[Joost van de Weijer]

Dear Info-childes members,

I was wondering if anyone could help me with the following:

I have collected a corpus of linguistic input to a prelingual infant which
includes language spoken not only directly to the infant but also to
others who were usually in the infant's environment, including an older
sister, the parents, and other caretakers.

I want to use my corpus to test the following hypothesis: assuming that
adults adapt their way of speaking to the child's linguistic level, I
expect that language to a prelingual infant is more abstract (i.e., it
contains relatively many words that cannot easily be related to the
concept they refer to) than language to a child who already has started to
produce or understand words.

In order to test this hypothesis, I made a basic classification of words,
dividing them into nouns, adjectives, adverbs, verbs, pronouns, proper
nouns, auxiliaries, determiners, and prepositions, because I thought that
the distribution of these categories would give me a preliminary
indication of how well the hypothesis is supported by the data. For
instance, perhaps the percentage of nouns is higher in language to the
older sister than to the infant because, in general, nouns are relatively
concrete compared to verbs or adjectives.

However, I am aware that within each category, there are differences too,
a noun like 'ball' being easier than 'side' for instance.  My question
therefore is: does anyone have a suggestion about how to make a
classification within the categories, or can anyone give me references of
comparable studies where I can find information on how to proceed?

Any responses are more than welcome,



Joost van de Weijer
Max Planck Institute for Psycholinguistics
PO Box 310
6500 AH  Nijmegen
The Netherlands
Tel: +31-(0)24-3521307
Fax: +31-(0)24-3521213
email: vdweijer at mpi.nl