Naslov
Implications of derivational transparency on the acquisition of lexicon

Autori
Palmović, Marijan ; Hržica, Gordana ; Mustapić, Maja

Vrsta, podvrsta i kategorija rada
Sažeci sa skupova, sažetak, znanstveni

Izvornik
10th International Congress for the Study of Child Language in Berlin : abstracts / Bittner, Dagmar ; Gagarina, Natalia - Berlin, 2005, 178-178

Skup
International Congress for the Study of Child Language in Berlin (10 ; 2005)

Mjesto i datum
Berlin, Njemačka, 25.07.2005. - 29.07.2005

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
derivational transparency; Croatian morphology; semantic bootstrapping

Sažetak
Derivational transparency is a prominent feature of Croatian morphology. If a single root is chosen in Croatian to derive a dozen of new words and if they are translated to, e.g. English, the translations will be words of different roots. For example, bol \'pain\' will produce bolnica \'hospital\', bolestan \'ill\', bolesnik \'patient\', bolni?arka \'nurse\' etc. This derivational productivity and transparency of derived meanings influence the course of lexical acquisition facilitating the acquisition of lexicon by providing a separate bootstrapping mechanism. It consists of language-internal information that provides semantic cues for recognizing grammatical form of a word. This mechanism differs from the semantic bootstrapping because no language-external information is involved, such as general cognitive notions for \'thing\' or \'action\' to enable the detection of nouns and verbs, as originally suggested by Pinker. Children take advantages of these lexical cues not only to detect meanings of derived words, but also to categorize them into appropriate lexical categories and deduce syntactic information. In this study two sets of data will be used to describe this language mechanism in more detail and provide theoretical account for it. First, a meta-analysis of cross-linguistic data will be done to show the differences in the acquisition of lexicon between Croatian and English. The amount of derived forms will be compared between the Croatian corpus in the CHILDES data bank and Brown\'s corpus (also included in the CHILDES). The biggest difference can be found in adjectives partly because possessive adjectives are early-developed mean of expressing possessiveness and partly because verbal adjectives are very frequent in Croatian. Overgeneralizations that involve derivations will be discussed, e.g. expressions in which children put a prefix or even a preposition onto a word to modify its meaning where a more analytical expression or different word should be used (e.g. *oko-rezati \'to cut around\' in expressions like *okorezati jabuku \'to cut around the apple\' instead of guliti jabuku \'to peel the apple\'). These overgeneralizations show how children choose the derived word where parts of its meaning are known rather than a new and unrelated word. The second set of data consists of a language test in which children are presented with word-selection and picture-selection tasks in which derivationally motivated words are offered together with control words of equal frequencies. Children tend to choose words that are derivationally motivated or understand their meaning better due to the lexical cues they can use. Learning task is reduced to adding an appropriate suffix based on the particular lexical category to which the new word belongs. Although language specific, the present data offer another perspective on bootstrapping due to the cues that are language-internal, although not of syntactic, but rather semantic nature. Since these cues are in fact lexical roots, this mechanism could be seen as \"lexical bootstrapping\".

Izvorni jezik
Engleski

Znanstvena područja
Filologija

POVEZANOST RADA