Applied Psycholinguistics 24 (2003), 453–489Printed in the United States of AmericaDOI: 10.1017.S0142716403000237The processing of ambiguoussentences by first and secondlanguage learners of EnglishCLAUDIA FELSER, LEAH ROBERTS, and THEODORE MARINISUniversity of EssexREBECCA GROSSUniversity of Du¨sseldorfADDRESS FOR CORRESPONDENCEClaudia Felser, Department of Language and Linguistics, University of Essex, Colchester CO43SQ, United Kingdom. E-mail: [email protected] study investigates the way adult second language (L2) learners of English resolve relativeclause attachment ambiguities in sentences such as The dean liked the secretary of the professorwho was reading a letter.Two groups of advanced L2 learners of English with Greek or Germanas their first language participated in a set of off-line and on-line tasks. The results indicate that theL2 learners do not process ambiguous sentences of this type in the same way as adult native speakersof English do. Although the learners’ disambiguation preferences were influenced by lexical–seman-tic properties of the preposition linking the two potential antecedent noun phrases (of vs. with), therewas no evidence that they applied any phrase structure–based ambiguity resolution strategies of thekind that have been claimed to influence sentence processing in monolingual adults. The L2 learn-ers’ performance also differs markedly from the results obtained from 6- to 7-year-old monolingualEnglish children in a parallel auditory study, in that the children’s attachment preferences were notaffected by the type of preposition at all. We argue that children, monolingual adults, and adult L2learners differ in the extent to which they are guided by phrase structure and lexical–semanticinformation during sentence processing.Our understanding of how mature readers or listeners process their native lan-guage in real time has increased considerably over the last couple of decades.Results from sentence processing studies using a range of different psycholin-guistic methods and techniques have shown that the adult parser is capable ofaccessing and rapidly integrating various types of structural and nonstructuralinformation during comprehension (see Gibson & Pearlmutter, 1998, for re-view). Comparatively little is known, by contrast, about the way language learn-ers process input from the target language. Instead, both first language (L1) andsecond language (L2) acquisition research to date has focused primarily on the 2003 Cambridge University Press 0142-7164/03 $12.00Applied Psycholinguistics 24:3 454Felser et al.: Processing of ambiguous sentencesdevelopment of linguistic competence (compare, for instance, the articles inRitchie & Bhatia, 1996, 1999).It is conceivable, though, that at least some of the differences in linguisticproficiency that have been observed between children and adults and betweenL2 learners and native speakers are due to differences in the way languagelearners and adult native speakers process the target language. As regards L1acquisition, it is possible that children’s linguistic development is constrainedby capacity limitations of the developing processing system (Adams & Gather-cole, 2000). Mature L2 learners, on the other hand, already have a fully devel-oped language processing system in place. Given that we know very little abouthow and to what extent this system is employed in L2 processing, though, itcould be the case that the lack of nativelike ultimate attainment in the L2 is atleast partially attributable to parsing problems rather than to an inability to ac-quire nativelike grammatical competence per se, as is often assumed (comparealso Juffs & Harrington, 1995; Kilborn, 1992; or VanPatten, 1996). Specifically,to the extent that there is crosslinguistic variation in L1 processing (see Cuetos,Mitchell, & Corley, 1996; Frazier & Rayner, 1988; Gibson, Pearlmutter,Canseco–Gonzalez, & Hickok, 1996; MacWhinney, 1997, 2002; and Mazuka &Lust, 1990; among others), L2 learners may transfer nontargetlike processingstrategies from their native language to the L2, which could be the crucial bar-rier to acquiring full nativelike performance in the L2. It is also possible thatsentence processing in the L2 differs in more fundamental ways from adult L1processing, for instance, in that L2 learners may have difficulty with the on-lineintegration of different information sources (compare, e.g., Kilborn, 1992). Inshort, answers to the question of whether language learners process the targetlanguage in the same way as adult native speakers do are likely to have impor-tant implications for theories of both L1 and L2 acquisition.The present study aims to compare the processing strategies used by advancedL2 learners of English to those employed by adult native speakers and childL1 learners of English. The main empirical questions to be addressed are thefollowing:• Do adult L2 learners employ the same sentence processing strategies as nativespeakers of the target language?• To what extent are language learners capable of using and integrating phrase-structure and lexical–semantic information during processing?• Is there any evidence that adult L2 learners transfer processing strategies fromtheir L1?In order to address the above questions, we have examined how advanced L2learners of English from different language backgrounds resolve relative clauseattachment ambiguities in two-site contexts, a phenomenon that has been exten-sively studied with monolingual speakers in the past and that is known to besubject to crosslinguistic variation. To explore the extent to which L1 and L2processing are similar (or dissimilar), we contrast the results from the L2 learn-ers with those from parallel studies with adult native speakers and English-speaking children that have recently been conducted by our research group.Applied Psycholinguistics 24:3 455Felser et al.: Processing of ambiguous sentencesThe remainder of this paper is organized as follows. The next section providesa brief overview of crosslinguistic differences in the processing of relativeclause attachment ambiguities, the phenomenon under investigation, followedby a summary of the results from existing studies of ambiguity resolution in L2processing in the next section. The following sections present the results fromfour experiments investigating relative clause attachment preferences in differ-ent groups of advanced L2 learners of English using both