Friday 8 September, 2017
Illusions of grammaticality in non-native language processing
Claudia Felser, University of Potsdam (Falta inscriure)
Speaker: Claudia Felser, University of Potsdam
Authors: Claudia Felser, University of Potsdam (Falta inscriure)
Grammatical illusions (GIs) may be reflected in erroneous positive judgements of ungrammatical sentences and/or in readers' or listeners' real-time processing patterns. In native speakers, GI effects can be triggered by the presence of an illusory licenser such as the plural noun cabinets in *The key to the cabinets were rusty, which may lead to processing facilitation relative to an equally ungrammatical control condition. Examining GIs can help reveal the nature of the linguistic constraints and memory mechanisms involved in parsing.
In illusory licensing contexts, GI effects are thought to reflect the parser's attempt to link a dependent element to a feature-matching (or partially matching) but grammatically inappropriate licenser in the absence of a matching legitimate licenser (Wagers et al., 2009). The erroneous acceptance of ungrammatical sentences might be considered a GI effect even in the absence of an illusory licenser, however. I will focus here on GIs that arise from errors in establishing grammatically licit relationships between dependent elements and their licensers. Barring cases in which incorrect judgements simply reflect a lack of grammatical knowledge, erroneous positive judgements may indicate insensitivity to grammatical cues during processing, an inability to carry out relevant feature-matching operations, or a failure to identify (or retrieve from memory) a dependent element's grammatical licenser. A broad view of GIs might also include the absence of grammaticality effects in processing tasks (e.g. Jiang, 2004, 2007; Keating, 2009; Sato & Felser, 2010) and cases of online parsing decisions that are not licensed by the grammar, such as the postulation of gaps inside syntactic islands (Boxell & Felser, 2017) or the retrieval of grammatically illicit antecedents for anaphors (Felser & Cunnings, 2012).
Both native and non-native speakers are susceptible to GIs in processing tasks, but depending on the type of phenomenon under investigation, the experimental task and other factors, non-native speakers may be more likely (e.g. Boxell & Felser, 2017; Felser & Cunnings, 2012), equally likely (e.g. Drummer & Felser, submitted; Lago & Felser, submitted) or less likely (e.g. Schlueter et al., 2017) than native speakers to show GI effects. Comparing real-time processing data from native and non-native speakers can be informative about the real-time status of linguistic operations, how these interact with processing constraints, and how these interactions may vary across populations. Recent findings from agreement attraction and filler-gap processing do not support the hypothesis that non-native speakers should generally be more prone to memory retrieval interference than native speakers (Cunnings, 2017). They do however indicate that L2 speakers have more difficulty than native speakers using morphosyntactic cues effectively during processing, and a greater reliance on semantic or discourse-level cues when establishing grammatical dependencies.
08:30 - 11:30 / Workshop 2 - “Grammatical processing in second language speakers: agreement, polarity, anaphora and other grammatical illusions”
Oral presentationThe processing of Negative Polarity Items in Turkish-German bilingual speakers
Speaker: Seckin Arslan, Universität Potsdam
Oral presentationA cue-based approach to long-distance anaphor binding in L2 Chinese
Speaker: Cecile De Cat, University of Leeds
Oral presentationOnline syntactic processing in native and non-native speakers of English: exploring quantitative and qualitative differences
Speaker: Carrie Ankerstein, Universität des Saarlandes