Hoekstra, E. and Slofstra, B. and Versloot, A.P. (2009) "Interacting mental lexicons / grammars of bilingual speakers: how Dutch influences Frisian".
|Microsoft PowerPoint (Presentation 'Morfologiedagen' Liege)|
Official URL: http://depot.knaw.nl/5605
Eric Hoekstra, Bouke Slofstra and Arjen Versloot, Fryske Akademy, KNAW Interacting mental lexicons / grammars of bilingual speakers: how Dutch influences Frisian The mental lexicons of bilingual speakers influence each other, as is well-known, but how this process takes place is less known. We argue on the basis of three case studies that both the frequency of Dutch words as well as their grammatical properties affect their Frisian equivalents, especially if the Frisian equivalents are phonologically similar to Dutch. Case 1. We present a statistical analysis of the role of the schwa in Frisian compounds. Compounds are formed from a left word and a right word: LEFT WORD + RIGHT WORD = COMPOUND brogge + tafel = broggetafel boterham tafel ontbijttafel mis(se) + boek = misboek mis boek misboek It will be shown that two factors influence the presence or absence of schwa in compounds: • whether or not the left word input to compounding has a Dutch equivalent that is phonologically similar • whether or not the left word has an optional schwa in Frisian or not. Case 2. Frisian possesses two native nominalising suffixes which compete with each other, -ENS and -HEID: batsk + ens = batskens * batskheid bot NOM = botheid wiis + heid = wiisheid * wizens tinkber + ens = tinkberens tinkber + heid = tinkberheid -HEID is supported by Dutch, but not -ENS. It turns out that -ENS is significantly more often used with adjectives that do not have a Dutch equivalent that is phonologically similar (batsk). Conversely, -heid is characteristically used with adjectives that do have an equivalent that is phonologically similar to Dutch (wiis). This feature interferes with a frequency component, where frequency has impact on the words that resemble a Dutch counterpart and share a similar meaning! High frequency supports the use of –heid. Case 3. It is well-known that verb clusters of three verbs may exhibit any order in so-called Interference Frisian (Koeneman & Postma 2006) , where Standard Frisian only allows a head-final order and Dutch a head-initial one: Standard Frisian: Omdat er dat dwaan wollen hie Standard Dutch: Omdat ie dat had willen doen Interference Frisian: All orders that are logically possible, though not equally frequent, Koeneman & Postma 2006:128. The word order facts follow from the assumption that Frisian verbs may choose either the word order checking specification of Standard Frisian (head-final) or that of Dutch (head-initial). This explains that every conceivable mix of head-final and head-initial is actually found. This could imply that grammar is an abstraction from the feature structures associated with lexical items. All in all, the results we obtained seem to support models making use of neural networks, and models in which constructions exist as independent entities. This view of bilingualism reveives further support from the fact that Frisian verbs in potential IPP contexts favour the use of -EN participles, whereas in non-IPP contexts, -EN and -D participles exhibit free variation.
|Deposited On:||21 Dec 2009 01:00|
|Last Modified:||22 Oct 2010 17:06|
Repository Staff Only: item control page