Upcoming Events

To date, my research has focused on the evolution of language-related cognitive abilities taking a comparative approach. Research projects investigating the speech perception abilities of a language-trained chimpanzee have shown similar performance as humans, providing evidence that speech perception abilities are rooted in general auditory mechanisms and most likely were present in a common ancestor of humans and chimpanzees. Additionally, the chimpanzee's performance has stressed the importance of early immersion in a speech-rich environment to facilitate speech perception. Results of experiments investigating the sequence learning abilities of rhesus macaques have shown rudimentary, rule-like, visual learning by the monkeys, providing evidence that these processing abilities may have been present in an early primate ancestor.

Colleen Balukas, Joe Bauman, and Cari Bogulski will talk about their experiences as Miccio Travel Award winners during the 2011-2012 award cycle.

Colleen Balukas Naomi Nagy at the University of Toronto and James Walker at York University. Joe Bauman visited Prof. Jessi Aaron at the University of Florida. Cari Bogulski visited Dr. Lee Osterhout at the University of Washington.

More about the Adele Miccio Travel Award can be found here.

Complement coercion, a form of enriched composition, occurs when two syntactically compatible, but semantically mismatching elements are combined in an expression (e.g., started + the book). The semantic mismatch triggers the coercion of one element from an entity into an event sense (e.g., started to read the book), resulting in processing costs when compared to control expressions. Some researchers believe that enriched composition is a primarily semantic process, and that it is the activation and selection of competing interpretations for the event (e.g., reading vs. writing the book) that cause processing delays. In addition, previous ERP work has shown that coerced nouns elicit an N400 response similar to that of semantically anomalous nouns (e.g., astonished the book; Baggio et al., 2009; Kuperberg et al., 2010). Coercion costs, then, should be sensitive to semantic manipulations inter- and intra-sententially. However, some behavioral research has shown that this may not be the case (Frisson & McElree, 2008; Traxler et al., 2005). I will present data from multiple experiments that, together, suggest that enriched composition is neither a purely semantically or syntactically driven process. In particular, these data demonstrate that event information in prior context can change the nature of the cost associated with coercion, but does not fully attenuate said cost.

In this talk, I explore the extent to which aspects of lexical representations and processing affect grammatical processing in late L2 learners. I will present data from two recent experiments on grammatical gender agreement (Experiment 1) and syntactic ambiguity resolution (Experiment 2).

In Experiment 1, twenty advanced to near-native L1 English speakers of German were tested on lexical gender assignment in production and grammatical gender agreement a visual-world eyetracking experiment. Performance on grammatical gender agreement in comprehension varied according to lexical gender assignment accuracy. Building on the lexical gender learning hypothesis by Grüter et al., (2012), I present a model to account for the dependencies between lexical and syntactic performance.

In Experiment 2, syntactic attachment preferences in 75 intermediate to advanced L1 German learners as well as 25 English native controls were tested. Unlike the natives who show syntactic attachment preferences, L2ers do not display clear preferences, which seems to point to the underuse of structural information in L2 parsing (e.g. Clahsen & Felser, 2006). However, once lexical processing factors – as independently assessed in a lexical decision task – are taken into account, robust and native-like structure-driven attachment preferences surface in adult L2ers, too.

I will highlight the role of lexical aspects on L2 grammatical performance and discuss the implications of these findings for L2 acquisition and L2 processing research.

Past research shows that lexical access is non-selective with respective to language, allowing cross-language interactions to occur (Dijkstra, 2005; Guo, Liu, Misra, & Kroll, 2011). A core question in bilingual research has been to understand the control mechanisms that allow bilinguals to select the language they intend to use. The study I will present aims to investigate the scope and the time course of inhibition in bilingual production. Language comprehension and production potentially differ in the way in which bilinguals achieve control of their two languages, e.g., in the time course of inhibition. In past research, we have shown that cross-language inhibition in comprehension seems to be relatively short-lived (Martín, Macizo, & Bajo, 2010). In contrast, studies of lexical production have shown that inhibition of the language not in use can be long lasting (e.g., Misra, Guo, Bobb, & Kroll, 2012), suggesting that there are multiple mechanisms of control. The present study explores the nature of the inhibitory mechanisms that underlie language selection in bilingual production and specifically whether there is evidence for both automatic and controlled selection processes. Relatively proficient Chinese-English bilinguals performed a picture naming task in language blocked or language mixed conditions (Guo et al., 2011) under simple naming conditions or while performing a concurrent and continuous updating task. Preliminary results show that the concurrent task affected performance differentially in the blocked and mixed conditions. I compare these results with the earlier comprehension studies and discuss the implications for models of cognitive control in production.

A rich history of work has led to several "truths" about bilingual language processing: 1) bilinguals activate words from all known languages even in single-language contexts (cross-language competition), 2) bilinguals divide daily exposure to each language, which leads to altered lexical entrenchment for language-unique words in a first and second language (L1 & L2), and 3) virtually everything about bilingual comprehension and production is modulated by L2 history, experience, and ability. While many theories of bilingual language processing generally accommodate these truths, they differ in the role ascribed to non-linguistic capacities such as cognitive control. Some models ascribe a secondary role to cognitive control (e.g., BIA+), arguing that they should only play a role following the earliest stages of word processing. Other models, such as Green's inhibitory control model and the bilingual advantages view posited by Bialystok and colleagues, ascribe a more central role to cognitive control in balancing within- and cross-language activation during bilingual production and comprehension.

In this talk, I highlight ongoing work from my laboratory that investigates whether individual differences in L2 ability and cognitive control relate to language comprehension and production processes among bilinguals. This work makes extensive use of eye-tracking, which has great temporal sensitivity and ecological validity for studying the earliest stages of both language comprehension and production. With respect to eye movement studies of sentence reading, we show that cross-language activation (interlingual homograph interference) is modulated by several factors including semantic bias of a sentence, L2 history, and, importantly, individual differences in cognitive control. Similarly, in eye movement studies of spoken language comprehension, we show that both within- and cross-language activation (generated by word-onset competition) is modulated by individual differences among bilinguals in L2 ability and cognitive control. With respect to language production, we show similar links between L2 ability and cognitive control when bilinguals produce L1 vs. L2 speech to describe a visual display while eye movements are monitored, and when they produce extended spontaneous speech in a monologue or dialogue context. Taken together, these studies suggest that cognitive control has much to do with bilingual language processing at the very earliest stages of comprehension and production, though much remains to be discovered about the specific kinds of cognitive control operations that are essential for specific bilingual language functions.

Word production in speaking culminates with the selection and articulation of word sounds. I will present a series of results - from neuroimaging (MEG) and acquired language deficits resulting from strokes - that shed light on the production of word sounds and their brain underpinnings. MEG correlates of picture naming show a fast activation of word sound processing, which starts about 150 ms after the picture presentation apparently simultaneously with the processing of the meaning of the depicted concepts. To the extent that these results suggest the concurrent activation of word sounds and meaning, they question the current view that, in picture naming, access to word semantics precedes access to word sounds. One of the implications of these results relates to the functional organization of brain mechanisms, as they suggest pathways directly connecting brain areas associated with the processing of visual object features and word sounds. On the other hand, the neuropsychological data provide strong support to the hypothesis of two distinct levels of processing in word-sound production, each associated with (at least partially) distinct brain mechanisms. At the phonological level, word sounds are encoded in an abstract form that does not specify context-specific features (e.g., phoneme length, aspiration) that are detailed at the following phonetic level of processing. I present results showing the each of these two levels of processing can be selectively damaged in conditions of acquired language deficits. Furthermore, results indicate that adjustments of word forms under the control of phonological grammar occur at the phonological level.

To date, my research has focused on the acquisition of relative clauses in monolingual typically developing (TD) children and monolingual individuals with neurodevelopmental language disorders, such as Down Syndrome (DS). A Relative Clause is a dependent clause that modifies a noun by making it more specific or adding additional information about it. Even though relative clauses emerge very early in the speech of TD children, some types of relatives are harder for them to comprehend and to produce. In this talk, I will focus on the production of the "harder" type of relative clauses, and I will compare TD children and individuals with DS, showing which solutions the two populations find to solve the problem of producing the syntactic structures.

One and the same word will sound somewhat different each time it is spoken, even if it is spoken by one and the same speaker. In fact, the actual pronunciation of words will often differ significantly from their "citation form" or "dictionary pronunciation". In this talk, I explore the respective roles of lexical retrieval times and of factors related to speech perception in pronunciation variation, both in conversational speech and in the lab. I argue that insights gained in psycholinguistic experiments, such as picture naming and lexical decision tasks, can shed light on lexical access and retrieval in conversational speech, and that current models of language production can usefully be extended to conversational speech.

Current research demonstrates that bilinguals are often advantaged relative to their monolingual counterparts on tasks that require cognitive control. A few past studies have also identified a bilingual advantage in the realm of word learning. However, these documented benefits of bilingualism are largely correlational, with little known about the underlying mechanisms that map language use to cognition and learning. One possibility is that all of the documented benefits of bilingualism reflect the effects of the constant mental juggling a bilingual at any age must exercise, as both of a bilingual's two languages appear to be active even when one language alone is required. Alternatively, different aspects of language use may map onto different types of cognitive consequences. The proposed research seeks to uncover the way that the use of multiple languages affects language learning as well as learning in other domains. The planned experiments will test the scope of the bilingual advantage in foreign language vocabulary learning by using electrophysiological measures that may provide a more sensitive index of the time course of early learning, compare the learning of linguistic and nonlinguistic information, and determine whether the bilingual advantage can be seen in the learning of a signed vs. spoken language. The goal of this set of experiments is to identify the cognitive mechanisms that underlie foreign vocabulary learning, and thus, to identify how learning a foreign language can be strategically enhanced.

A bilingual learner faces the task of learning and differentiating two languages. Recent research suggests that constantly switching between and inhibiting languages results in a language learner with distinct differences in comparison to a monolingual. For example, bilinguals show advantages in both linguistic and non-linguistic tasks measuring cognitive control (i.e., task switching, inhibition, attentional control). Recent work in statistical learning has begun to explore how learners detect and acquire multiple linguistic systems. Relatively few studies, however, have examined possible differences in the statistical learning mechanism or outcomes as a result of a learner's bilingualism. In a series of statistical learning tasks (i.e., speech segmentation and cross-situational word learning), we compared the performance of English monolinguals from Penn State University and Chinese-English bilinguals from Beijing Normal University. Preliminary results show faster and more robust learning for bilinguals in comparison to monolinguals, as well as differential effects of L1 identity on learning outcomes. These results suggest an advantage for bilinguals in statistical learning tasks, as well as an influence of L1 patterns on the learning of additional languages in sequential and unbalanced bilinguals.

In the standard neogrammarian view a distinction is made between regular, biomechanically-induced, sound change and psychologically-based analogy. In a sense, however, all sound change has a psychological aspect, even when its origin is in biomechanics, since at some point phonological recategorization is required for sound change to take place (e.g. /p/ > /b/). In Labovian sociolinguistic research a distinction is also made between change from below and change from above related to speakers' awareness.

In this presentation I will consider the role of phonological awareness in regular sound change drawing from my recent acoustic research on intervocalic consonant lenition in a number of languages (including Spanish, Italian and Basque). I will argue that, at an initial stage, lenition applies as the neogrammarians envisioned: across morphological boundaries and without regard to lexical identity. At this initial stage the process may be below speakers' consciousness, and yet may operate as a conventionalized reductive process in the speech community, beyond biomechanical reduction. A number of factors may cause awareness of the phenomenon and its phonologization. It is at this stage that word- and morpheme-boundaries start to matter as conditioning environments and we also find lexical effects.

This research has also revealed the existence of important individual differences in phenomena such as intervocalic consonant voicing, correlated in part with the sex of the speaker. I will discuss the possible eventual conventionalization of sociolinguistic variation from biomechanic biases in lenition processes perhaps through the social construction of these individual differences in speech.

The work I will be presenting includes data collected during my PIRE-sponsored visit to Dr. Teresa Bajo's lab at the Universidad de Granada, Spain. Verb subcategorization bias, often just called verb bias, is a usage-based property assigned to verbs based on the frequency with which that verb occurs with any of its allowable subcategorization frames. The most frequently co-occurring subcategorization frame, often based on corpus analyses of naturalistic data, is considered to be that verb’s bias (e.g., Spanish: Dietrich & Balukas, 2012; English: Gahl, Jurafsky & Roland, 2004). Recent research with monolingual speakers (e.g. Wilson and Garnsey, 2009) as well as proficient second language (L2) speakers (e.g., Dussias & Cramer Scaltz, 2008) has demonstrated that verb bias guides the initial selection of a structural analysis during online processing. Research with bilingual populations has shown that bilingual lexical access is non-selective (e.g. Schwartz, Kroll & Diaz, 2007), and that translation equivalents of verbs in different languages do not necessarily share the same bias (Dussias, Marful, Bajo & Gerfen, 2011). In light of these discoveries, the goal of the present study is to investigate whether verb bias information from Spanish (the native language, L1) is activated and used during sentence processing in English (L2) by examining how verbs with same and different biases in a bilingual's two languages impact initial syntactic analysis. Given that verb bias is based on usage frequency information, which presumably is developed through language experience, we furthermore investigate the role that immersion in the second language has on a bilingual’s processing strategies.

Numerous models of second language (L2) acquisition have been proposed, from either a linguistic (Clahsen & Felser, 2005; Herschenson, 2000; Schwartz, 1998;) or neurolinguistic perspective (McLaughlin et al., 2010; Osterhout et al., 2008; Steinhauer, White & Drury, 2009; Ullman, 2001). These models propose contrasting views as concerns the role of the native language in adult L2acquisition, the ultimate level of attainment that can be achieved, whether distinct neurophysiological responses are indicative of distinct stages of L2 processing capacity, and, indeed, whether the cortical areas involved in L2 processing are the same as in native language processing. In the present talk, I will present an overview of these models and their predictions as concerns online syntactic processing. Data from various ERP and eye-movement experiments run in my laboratory will be presented which, overall, question the idea of any strict series of neurophysiological responses linked to levels of L2 competence (Carrasco & Frenck-Mestre, in prep; Foucart & Frenck-Mestre, 2011, 2012), highlight the importance of using complementary methods to capture processing capacity (Foucart & Frenck-Mestre, 2012) and pinpoint protracted areas of processing difficulty and how they relate to the convergence of grammatical features across the L1 and L2 (Carrasco & Frenck-Mestre, in prep; Foucart & Frenck-Mestre, 2011, 2012).

The ability to suppress irrelevant information or stop prepotent responses is crucial for the efficient processing and interaction with our environment. The so-called mechanism of inhibition has been observed to play an important role in different domains such as attention, memory and, more recently, in language. One focus of interest is whether inhibition is an unitary mechanism acting on different representations (e.g. motor, mnemonic, linguistic, etc.) and under different tasks (an unitary view of inhibition) or whether there are multiple inhibitory processes. Those that support this last view consider that inhibition could be separated in terms of its dependence or independence on control processes (e.g. Nigg, 2000). This debate has promoted research from different approaches (developmental, neuroimaging, or clinical, for example) that support both views. Here we provide behavioral and neurophysiological data that favor an unitary view on inhibition as a control mechanism over memory and language.

Learning grammatical gender in a second language is a particularly difficult task for L2 learners. Much of this difficulty can be attributed to a lack of understanding of what role gender plays in the language. In the first of two experiments exploring grammatical gender, first-semester students were given in-class treatments based on the principles of Processing Instruction (VanPatten, 2004) in order to highlight the meaning associated with grammatical gender in German. We wanted to test whether, if given tasks that require the participant to pay attention to the gender in order to correctly solve a problem, L2 learners can begin to better acquire the German gender system. Even L2 speakers who can learn and use the gender system in German and who achieve high proficiency in the language in general still face challenges in terms of gender processing. The predictive use of grammatical gender in the processing of a language has been shown to differ greatly between L1 and L2 speakers (e.g., Scherag, Demuth, Rösler, Neville, & Röder, 2004; Lew-Williams & Fernald, 2010). Even speakers who have lived in an immersion environment for a long period of time and speak at an advanced level tend to ignore gender cues during processing (Hopp, 2011). Is the L2 speaker's inability to fully use gender in the same predictive way as a native speaker the result of learning the language too late in life or simply a matter of needing additional processing time to achieve the same result? In the second experiment we designed a task which first gave participants a gender prime followed by a sentence presented using RSVP. This simple sentence included an adjective between the definite article and the final picture in order to increase processing time. By measuring RT data between L1 and L2 speakers, it was determined that many L2 speakers were indeed able to use gender in a more native-like manner.

Most people associate Sigmund Freud with the assertion that speech errors reveal repressed thoughts, a claim that does not have a great deal of support. I will introduce some other things that Freud said about slips, showing that these, in contrast to the repression notion, do fit well with modern theories of language production. I will illustrate using an interactive two-step theory of lexical access during production, which has been used to understand aphasic speech error patterns.

A fundamental characteristic of language is that it provides multiple ways to express the same ideas, and therefore speaking presents the challenge of choosing between competing alternatives. Bilinguals provide a unique source of evidence about how speakers gain control over these selection challenges, given that they often face direct competition between languages. Current research suggests that bilinguals manage this competition with domain-general mechanisms of cognitive control. By implication, proposals that the language system may be equipped with its own specialized processing mechanisms are rejected. I will present a series of studies that question this basic claim by demonstrating that bilinguals with impairments in executive control (due to aging and Alzheimer's disease) exhibit relatively intact ability to do what bilinguals do best. This dissociation invites a psycholinguistic model that is equipped with at least some domain-specific control mechanisms, and that does not attribute all the consequences of bilingualism to mental juggling of two languages. These data also provide unique insights about why retrieval sometimes fails when people speak, and has practical applications for diagnosis of cognitive impairment in an increasingly multi-lingual society.

There is ample evidence that advanced second-language learners activate lexical information in both languages even in contexts where only one language is relevant. It is still unclear to what extent syntactic properties, such as word order, are activated in one language while processing the other. To test this, we compared native English speakers with native Dutch, advanced second-language learners of English. Tasks included a self-paced reading task, manipulating agreement and word order so as to create cross-linguistic ambiguities for the second-language learners. L2 learners showed less sensitivity to grammaticality manipulations than native English speakers. Effects of native language word order were not found during on-line reading, but could be observed in performance on an end-of-sentence statement verification task. Theoretical implications of these findings will be discussed. In addition, the effects of proficiency and cognitive control will be considered.

One of the main points I make in this talk is that now that we have learned a fair amount about heritage languages time has come for linguists to learn from them about the overall design of natural language. Both linguistic theorizing and experimental studies of language development rest heavily on the notion of the adult, perhaps linguistically stable, native speaker. Native speaker competence and use are typically the result of normal first language acquisition in a predominant monolingual environment, with optimal and continuous exposure to the language. In this talk, I discuss the case of heritage speakers, i.e., bilingual speakers of an ethnic or immigrant minority language whose first language does not typically reach native-like attainment in adulthood. I present an overview of heritage speakers’ linguistic system and discuss several competing factors that shape this system in adulthood. The examination of the linguistic knowledge of heritage speakers allows us to question long-held ideas about the stability of language before the so called critical period for language development, and the nature of the linguistic system developing under reduced input conditions.

Sara Carter - Processing of English Verb Bias in the Spanish L2 Immersion Environment

Sarah Fairchild - Determiner-noun codeswitching in Welsh-English bilinguals

Emma Hance - Does script influence novel word learning? A comparison of same-script and different-script bilinguals

Thomas Holt - Categorical Representation in Chinese Monolinguals

Melissa Magro - /s/ lenition in the speech of Spanish-speaking children from Granada

Jesse Martz - Overhearing a second language abroad as an adult: Learning with no intention

Elizabeth Mormer - Production of subject-verb agreement in English by Swedish-English bilinguals

Clair Pelella - The comprehension of codeswitches among speakers who don't codeswitch

Emily Sabo - Verb-bias and plausibility: Do L2 speakers use these sentential cues in the same way that L1 speakers do?

Previous empirical research has demonstrated the ability of infant and adult learners to track the statistical regularities in a novel language, relying on these patterns to discover underlying structures in the input. In speech segmentation specifically, learners are sensitive to transitional probabilities between phonetic units (such as syllables), identifying words within which transitional probabilities are relatively high and between which transitional probabilities are low. Recent studies suggest that this abi containing partially overlapping phonetic inventories unless an explicit cue to the change in input is provided. In the first experiment, we replicate the results of Gebhart et al (2009), demonstrating that learners fail to segment a second language when they are exposed to two different input streams for 5.5 minutes each sequentially. In the successive experiments, we attempt to reconcile the results of Experiment 1 with a previous study which indicated that explicit language cues were not always necessary to parse both languges (Weiss, Gerfen, & Mitchel, 2009). We vary language exposure (duration) and frequency of language switching, revealing that under certain combinations of these parameters, both languages may be learned in the absence of explicit language cues. Finally, we explore some preliminary data indicating that the effects of these training conditions are not uniform between monolingual and bilingual adults.

The study examined whether the orthographic transparency of bilinguals’ first (shallow L1 – Spanish; deep L1 – Chinese) language has an impact on the reading strategy they adopt in an orthographically deep second (L2 – English) language. Highly proficient Spanish-English and Chinese-English bilinguals and English monolingual controls made rhyme judgments of visually presented English words while behavioral and EEG measures were recorded. The spelling-sound consistency and orthographic similarity of semantically unrelated rhyming and non-rhyming prime-target pairs were varied systematically. To manipulate consistency, graphemically dissimilar primes and targets that either matched or did not match in consistency were compared in both the rhyming (consistent/consistent: WHITE-FIGHT versus inconsistent/consistent: HEIGHT-FIGHT) and non-rhyming conditions (consistent/inconsistent: SCALE-LEAK versus inconsistent/inconsistent: WORK-LEAK). Orthographic similarity was manipulated by comparing pairs that matched in consistency, but were either graphemically dissimilar (WHITE-FIGHT; WORK-LEAK) or graphemically similar (RIGHT-FIGHT; STEAK-LEAK). Results suggest that bilinguals with a shallow L1 orthography may treat words in a deep L2 orthography as equally inconsistent in the absence of converging cues from orthography and phonology, whereas bilinguals with a deep L1 orthography may treat words in a deep L2 orthography as equally consistent.

In the nineteenth century, the Neogrammarians hypothesized that sound change is regular and attributed exceptions to sound changes to factors such as analogy and dialect mixture. The explanation of both regularity and exceptionality in phonological change remains one of the primary challenges in historical linguistics regardless of theoretical framework (see Kiparsky, 1994; Labov, 1994, 2011; Bybee, 2003; Blevins & Wedel, 2009; Bermudez-Otero, 2010). This paper investigates the lengthening of stressed short vowels in open syllables in Middle Dutch, Middle English and Middle High German.

Of particular interest is the regularity of Open Syllable Lengthening in Middle Dutch versus the large number of exceptions in Middle English and Middle High German. Previous work has attributed the exceptions in Middle English and Middle High German to intervening sound changes, analogy or dialect mixture. I will argue that Open Syllable Lengthening (OSL) is motivated by speech perception and a reparsing of tonic short vowels as long by listeners (Kavitskaya, 2002). Unlike many regular sound changes that have an articulatory basis, OSL obeys structure preservation and is not reductive. In this regard, OSL is similar to so-called sporadic sound changes, such as metathesis and dissimilation, which have long been recognized as irregular and admitting exceptions. The regularity of OSL in Middle Dutch is accounted for by the loss of contrastive vowel length and the reanalysis of vowel length as a predictable feature of lexical stress.

Learning a second language as an adult is arguably one of the most difficult learning tasks that one can undertake. Yet, it is of great importance in our increasingly global society. In order to fully understand second language development in adults, we must understand not only the processes involved in acquisition but also how these processes are affected by external and internal factors. In this talk, I report the results from a series of four studies aimed at elucidating the role of both external and internal factors in the developing (neuro) cognitive underpinnings of adult second language acquisition of grammatical structures. In each of the studies, adults learned an artificial second language that was modeled after and consistent with natural language. Participants learned to speak and comprehend the second language in order to refer to the pieces and moves of a made-up chess-like computer game. In the first two studies, the effect of an external factor—the condition under which the artificial language was learned—is examined in light of learners' performance as well as their neurocognitive processes (as revealed by event-related potentials). In a subsequent study, the role of internal factors is addressed. Specifically, this experiment explores whether individual differences in cognition can predict learners' development in the artificial second language. performance measures and measures of the neural representation of the acquired artificial language (as revealed by functional magnetic resonance imaging). The final study explores the potential interaction between external and internal factors by examining how learners' individual differences in cognition affect their performance under different training conditions. Results from these studies suggest that adult second language learners can achieve and retain processes similar to that of native speakers, though only when they attain high proficiency. Furthermore, attainment of high levels of proficiency and native-like processing appear to depend on certain factors, including linguistic structure, the condition under which the language is learned, and individual differences in cognition. The implications of these results will be considered in the context of both theoretical and applied questions related to successful second language acquisition, and future research directions, including longitudinal research with natural languages, will be discussed.

Though I am not a language scientist, I have rubbed shoulders with some of them over the years. Despite my ignorance of this field, but encouraged by the open-mindedness and welcoming attitude I see among you, I'd like to offer the following four questions and answers for your consideration:

1. Are you actually studying language per se? Not necessarily.

2. Why is there more than one language? To promote incomprehension.

3. Would your research benefit from the statistical technique of bootstrapping? Quite possibly.

4. What limits foreign language learning? Incomplete goals.

The overarching purpose of this study was to evaluate the effects of the research-based reading program, RAVE-O, for children (N=60; ages 6;10-9;0) in a local community within a summer school setting. The working hypothesis underlying this intervention was that the intensive 4-week summer reading intervention described in this talk would reduce the achievement gap by focusing instruction on the componential processes involved in fluent reading and reading comprehension. Three treatment conditions were used to evaluate the effectiveness of various repeated reading techniques (repeated reading alone, listening during repeated reading, and accelerated repeated reading) on subsequent tests of fluency and reading comprehension. The treatment conditions were evaluated according to several demographic variables known to moderate the achievement gap: socio-economic status, language background, and reading disability. Results confirm the efficacy of multi-componential reading interventions in reducing the summer achievement gap. Importantly, results indicate that all the varied applications of the treatment conditions differentiated typical readers from their RD peers with moderators having differential impacts depending on specific treatment condition. In other words, learner characteristics appear to influence responses to specific reading instruction practices.

Bilingual speakers—even those who frequently engage in code-switching—are normally aware of what language they are speaking at any given time, and can correctly identify exemplars of each language. This is not always the case for bi-dialectal speakers, including a national standard language and a regional dialect (e.g. standard Italian vs. regional “dialects,” also Spanish-Portuguese in northern Uruguay). There is, however, no widely accepted consensus on the degree of morphosyntactic similarity between genealogically related and partially cognate systems that mark the psycholinguistic threshold of identification as distinct languages. One possible testing ground involves heavily restructured languages such as creoles in contact with their historical lexifier languages. The present study presents data from the creole language Palenquero, spoken in San Basilio de Palenque (Colombia), which has been in contact with its principal lexifier, Spanish since the formation of the community in the late 17th century. The Palenquero language, known locally as Lengua ri Palengue (LP), exhibits a number of key grammatical features found in no variety of Spanish. Mutual intelligibility between Spanish and LP is in general quite low; monolingual Spanish speakers may recognize individual words in LP, but cannot accurately parse LP syntactic structures, which include: absence of grammatical gender, marking of nominal plural with the preposed particle ma rather than the (multiply-agreeing) suffix /-s/, invariant verbs with preverbal tense-mood-aspect particles, negation by clause-final nu, absence of definite articles, a single set of obligatorily overt pronouns (all different from Spanish), marking possession by postposing the possessor. Most of the differences between Spanish and LP are categorical and binary; it is therefore not unreasonable to assume that Palenqueros psycholinguistically partition Spanish and LP according to such parameters, that they are able to identify given configurations as belonging to either Spanish or LP, and that utterances containing both quintessentially LP and uniquely Spanish structures will be acknowledged as mixed. Palenqueros do not exhibit a “post-creole continuum,” i.e. a systematic cline of intermediate variants spanning the linguistic distance between the creole language and its lexifier language. These facts notwithstanding, linguists who have studied contemporary LP have noted the frequent introduction of indisputably Spanish elements, ranging from individual items such as conjugated verbs or preverbal clitics to more complex morphosyntactic constructions. Opinions as to the nature of this apparent mixing—rarely substantiated by empirical data—include decreolization, language attrition, code-switching, interference from Spanish, performance errors, and the possibility that such configurations have been an integral part of LP since its origins. Equally difficult to extract from available studies are Palenqueros’ implicit and explicit notions of “canonical” LP as well as their awareness of putative deviations from any loci of inter-speaker acceptance. The present study is based on experiments conducted in San Basilio de Palenque, using stimuli extracted from natural speech samples. The stimuli included utterances entirely in Spanish, entirely in LP (as described by native speakers, e.g. Pérez Tejedor 2004, Simarra Obeso et al. 2008, Simarra Reyes et al. 2008), and containing what might be considered Spanish-LP morphosyntactic mixing. In the first experiment respondents were asked to classify stimuli as Spanish, LP or mixed. All-Spanish and all-LP utterances were almost always identified accurately but there was considerable diversity in reactions to putatively mixed stimuli. Responses were subjected to a variationist analysis to determine the factors that influence language classification. Language-specific pronouns, presence or absence of feminine gender marking, and speaker status (young, older traditional, LP language teacher) were the most significant factors, while conjugated verbs, preverbal clitics, definite articles, and other “Spanish-like” elements were not significant predictors of “mixed” responses. In a second experiment the same participants close-shadowed all-Spanish, all-LP, and nominally mixed stimuli. The rationale of such tasks is that “when listeners hear a sentence that exceeds the capacity of their short-term memory, they will pass it through their own grammar before repeating it” (Gullberg et al. 2009: 34). Previous work (e.g. Marslen-Wilson 1973, Vinther 2002) has shown that in sentence repetition tasks, respondents’ errors frequently reflect their own grammars, i.e. what they would have said instead of what was actually said. No language switches or other grammatical alterations were made for all-Spanish and all-LP stimuli, but there were numerous spontaneous “corrections” of mixed utterances, almost always resulting in all-LP combinations. There was also a strong correlation between “mixed” judgments and spontaneous correction by the same respondents during shadowing. The overall results suggest that code-switching as commonly defined is not explicitly accepted by Palenqueros. They also demonstrate an asymmetry between perception and production: “grammars” and “languages” are not psycholinguistically coterminous for LP-Spanish bilinguals. Despite apparently clear-cut distinctions between Spanish and LP, grammatically-defined boundaries have been partially supplanted by a more amorphous duality based on a combination of key lexical items, phonotactic profiles, and acknowledgment of known speakers as “true” Palenqueros. The linguistic situation of San Basilio de Palenque demonstrates the challenges for scholars and teachers seeking to define and delimit bilingualism in the absence of community-wide literacy, accepted canonical standards, and a rapidly evolving metalinguistic awareness.

In this talk I will address syntactic processing and ERPs in both L1 and L2 (in 2 parts). The first part will demonstrate that (and why) some of the most influential ERP papers on syntax in native speakers have major methodological problems, which affect in particular early ERP effects that have been taken to support 'syntax-first' models (such as ELANs in Friederici's 2002 model; for discussion see Steinhauer & Drury, in press).

Interestingly, the absence of such early ERP components in L2 learners has been interpreted as strong evidence for critical periods in L2 syntax acquisition (e.g., Weber-Fox & Neville, 1996, and Hahne & Friederici, 2001). However, the few ERP studies that have argued *against* the critical period hypothesis typically relied on flawed data as well (e.g., Rossi et al., 2006).

Therefore, in part 2 of my talk I will discuss various problems in L2 research and present some ERP data in this domain that my students and I collected. I will argue that ERP evidence for critical periods in L2 morphosyntax is weak and show how L2 ERPs systematically change with increasing L2 proficiency (including 'native-like' ERP profiles at very high levels of L2 proficiency; e.g., Steinhauer et al., 2009). These changes are modulated by factors such as L1 background (co-activation and transfer) and the type of L1 exposure (classroom vs. immersion; e.g., Morgan-Short et al., in press).

The literature agrees: the acquisition of the Spanish subjunctive mood is difficult for second language (L2) learners of Spanish. Although the acquisition of the subjunctive has been studied from a variety of perspectives (Collentine, 1997; Farley, 2004; Gudmestad, 2006; Isabelli, 2007; Lubber’s Quesada, 1998), no studies to date have examined whether the knowledge that advanced speakers of Spanish have acquired about the subjunctive is used during on-line processing. In this presentation, we examine whether highly proficient L2 speakers of Spanish can use the information they have learned about the Spanish subjunctive – even if that information is below the native speaker mark – to predict its occurrence in subordinate clauses. Data from L1 Spanish monolinguals living in Spain, and from Spanish-English and English-Spanish speakers living in the United States were collected. Analyses show that the monolinguals and the Spanish-English bilinguals are able to correctly predict verbal mood in subordinate clauses, replicating previous findings in the literature (e.g., Demestre & García-Albea, 2004). However, the English-Spanish group living in the U.S. did not show the same sensitivity.

Research suggests that L2 learners show initial difficulty with learning how to interpret the pre-verbal clitic structures in Spanish (e.g. Liceras, 1985). Whether this difficulty in processing OVS structures is based on a learner’s L1 or proficiency level has just begun to be examined (e.g. Seibert Hanson & Sagarra, 2010). In addition, Havik et al. (2009), comparing English- and German-Dutch learners, found a similar difficulty with OV interpretation, highlighting working memory capacity (WMC) as a mitigating factor. Romanian, like Spanish, allows for preverbal direct object clitic sentence structures (e.g. O caută băiatul, “Her-OBJ looks for the boy-SUBJ”). If the L1 is the determining factor in early success with preverbal clitics, then L1 Romanian speakers may be more successful with such structures in Spanish than L1 English speakers, who do not possess this structure in their L1. Contrariwise, if there is some level of proficiency that must be reached universally to properly process preverbal clitic structures, there should be no differences between learners based on their L1. Additionally, variation within language may be due to WMC.

To test this, 65 L1 English learners of Spanish and 71 L1 Romanian learners of Spanish, matched for proficiency, and 35 Spanish monolinguals completed a WM task and a task in which they heard sentences with preverbal clitics in Spanish (e.g. Lo besa la niña, “Him-OBJ kisses the girl-SUBJ”) and choose from four pictures the one that was most accurately described. One-way ANOVAs revealed significant differences among learner participants based on proficiency level, but not on L1. However, the OVS accuracy means for the Romanian-Spanish learners were greater than the means for the English-Spanish learners, indicating that other factors are involved. Results from further analyses will be revealed, showing the role WMC plays in the variance among the participants. The present data suggest that working memory capacity plays an important part in determining the accuracy of interpretation of OVS structures at early stages of acquisition, and that language experience influences when L1 transfer takes place during language acquisition.

The acquisition and resulting grammar of heritage languages represent significant challenges to current theories of learning/acquisition and associated grammatical formalisms. In our talk, we present new results from ongoing work by our team.

To frame the project, we first revisit recent proposals by Montrul (2002, 2004, 2008, 2009) and Polinsky (1997, 2006, 2008) who draw a clear distinction between incomplete acquisition, i.e., the situation of early/sequential bilinguals, andL1 attrition, i.e., the situation of the loss of certain performance-oriented skills that do not affect the core elements of the competence grammar. We note some complexities and difficulties with the notion of incomplete acquisition, and are working to develop a straightforward way of addressing heritage language acquisition from the standpoint of feature activation, and with a focus on variation and change within the community. From there, we present findings on two morphosyntactic phenomena in heritage German grammar, drawing data from a wide range of settings from free conversation to grammaticality judgments.

First, we take a closer, detailed look at the passive voice construction inventory in Moundridge German — a moribund German-language speech enclave (Eastern Palatinate) with ca. 40-50 remaining speakers of various degrees of fluency in the dialect. Data from Moundridge German (Putnam & Salmons, under review) shows a reduction in possible passive voice constructions in the final stage of the MG-grammar. These findings make an interesting contribution to theoretical debates regarding the modeling of the loss of (syntactic) grammatical categories. Based on our findings, we show how these data suggest that a neutralization approach is preferable to null parse strategies in accounting for syntactic ineffability (Legendre et al. 2006, Legendre 2010).

Second, we explore the presence or absence of parasitic gap constructions in a set of heritage German dialects spoken near Sheboygan, Wisconsin, descendants of mid-19th century German who still speak local varieties. Parasitic gap or ‘multiple gap’ constructions represent a interesting domain of investigation here on two grounds: (1) There is a sharp contrast between the two grammars in contact. English allows parasitic gaps in a multitude of environments, whereas it is generally argued that German lacks these constructions altogether or licenses them in only very specific contexts (e.g., Parker (1999), Kathol (2001), Chocano & Putnam (in press)). (2) Null elements (such as gaps) have been claimed to be rarely attested in heritage language grammars (cf. Polinsky & Kagan 2007). Our pilot data challenges the expected lack of null elements/gapping structures in heritage grammars: Our speaker show a range of patterns, from close adherence to German-like avoidance of gaps to various degrees of gapping. Results overall correlate with proposals for the complexity of parasitic gap constructions.

In both case studies, we can identify a cline from full L1/native-like grammars to considerably reduced grammars. They also reveal complex patterns of interaction between German-like patterns and English-like patterns in the heritage grammar.

According to received wisdom, other-language words are introduced into recipient-language discourse by a bilingual speaker, gain in frequency, become linguistically integrated, diffuse throughout a community of bilingual and eventually monolingual speakers, and finally achieve dictionary attestation and native status. But how just how do they get from there to here?

Previous historical research on lexical borrowing deals with the product, i.e. attested loanwords, and the few empirical synchronic studies that treat the process as it occurs spontaneously in the bilingual community are necessarily silent on the diachronic trajectory that borrowed forms follow. In this paper we address these issues by tracking the evolution of English-origin material in a unique data set on Quebec French collected over a real-time period of 61 years, and panning nearly a century and a half in apparent time. From these corpora, we extracted close to 19,000 tokens of lone English-origin items and 2,000 multiword fragments of English.

Through detailed quantitative analyses of those items that persisted over the entire duration, and others that were short-lived, we address three widespread beliefs about the processes underlying code-switching and borrowing:

1) Other-language incorporations are introduced as nonce forms and gradually increase in frequency and diffusion
2) Other-language incorporations are introduced in donor-language phonological, morphological and syntactic form, but they (or some subset thereof) are gradually integrated into recipient-language structure, in tandem with increases in frequency and diffusion
3) At least in the earliest stages, and possibly throughout, code-switching cannot be distinguished from borrowing.

Results provide little evidence in favor of any of these hypotheses. Surprisingly few other-language items persist, even over the relatively brief period of time studied here, let alone increase in frequency or diffusion. Linguistic integration is abrupt, not gradual. Speakers all but categorically integrate lone other-language items at first mention, while never treating multi-word fragments of the other language in this way. This in and of itself is evidence that the two classes of other-language item—single-word vs. multi-word—are demonstrably different. But they also differ wildly by their word-class and grammatical constitution, their overall frequency of occurrence, and the relative propensity of a given speaker and a given community to use one rather than the other. We explore the implications of these results for understanding the processes by which other-language incorporations achieve the status of native items, and their consequences for theories of code-switching and borrowing.

Habitual aspect in Palenquero, a Spanish-based creole spoken in San Basilio de Palenque, Colombia, has been described as being expressed by the preverbal marker asé (from Spanish hacer ‘do’), as in the example (1) below. (Schwegler 1992: 224, Schwegler & Morton 2003, cf. Davis 1997: 27-30).

(1) Ahora nu. Majaná asé salí cu sei u siete u ocho majaná. Today NEG. Kids HAB go out with six or seven or eight kid. 'Not these days. Kids go out with six or seven or eight kids'.

This pilot study analyzes variable use of asé, using typological insights from grammaticalization theory (Bybee et al. 1994) and the variationist method (Labov 1966) to uncover distributional patterns. The data for this study were taken from sociolinguistic interviews and conversations with 10 speakers recorded* during July, 2010 and May, 2011 (thanks to funding from the Center for Language Science) in San Basilio de Palenque, including male and female participants, ranging from high-school age to older speakers.

The distributions suggest that tense-aspect marking in Palenquero is not captured by a one-to-one mapping of form and meaning. Habitual is expressed by both asé and ‘zero’, although it has been suggested that in creole languages zero marking indicates present tense on stative verbs and past tense for non-statives (Bickerton 1975, 1981, 1984). Asé, on the other hand, is more closed associated with frequentative meaning than with habitual. This is consonant with the suggestion in Bybee et al. (1994) that habitual meaning develops out of frequentative meaning. These preliminary data suggest that grammaticalization theory can neatly account for the distributional patterns seen in the synchronic creole data. Other questions are raised, though, as to whether the monogenetic view of grammaticalization should be assumed in working with creole languages, and if so, what is its place in creole studies. Based on these questions, I will also discuss directions for further research.

*Thanks to Colleen Balukas, Amelia Dietrich, and in particular, Dr. John Lipski for generously providing some of their recordings.

One of the fundamental processes of speech perception is a contextual normalization process in which segments are "parsed" so that the effects of coarticulation are reduced or eliminated. For example, when consonants are said in sequential order (e.g. the [ld] in "tall dot" or the [rg] in "tar got") the tongue positions for the consonants interact with each other. This "coarticulation" is undone in speech perception by a process that is called compensation for coarticulation. The basis of this process is a source of much controversy in the speech perception literature.

I studied the compensation for coarticulation process in two ways. The first set of experiments examines the role of top-down expectations, finding that the compensation effect is produced when people think they hear the context, whether the context is present or not. This dissociation of the context effect from any acoustic stimulus parameter indicates that at least a portion of the compensation effect is driven by expectations. The second set of experiments examines the role of articulatory detail in the compensation effect, finding that the compensation effect is driven at least partly by detection of particular articulation patterns. This set of experiments looked at perception in context of the "retroflex" and "bunched" variants of English "r" and found that this low-level articulatory parameter is crucial for the compensation effect. The overall picture that emerges is one of listeners who make use of fine-grained articulatory expectations during speech perception.

Can certain sound patterns and lexical properties evoke the perception of gay-sounding male speech (GSMS)? Further, are the sound patterns and lexical properties associated with GSMS invariant across languages? Previous naturalistic research with English monolinguals has demonstrated that listeners are able to accurately determine sexual orientation from speech alone, although a thorough understanding of the acoustic and lexical correlates of GSMS remains elusive. Most studies of the stereotype to date have relied on explicit comparisons of recorded speech from self-identified samples of gay and straight speakers; this research has identified a number of potential acoustic correlates, among them longer sibilant duration, higher peak sibilant frequency, and greater pitch range. Here we use an experimental approach to investigate how the manipulation of a single acoustic cue, sibilant duration, and several lexical properties (e.g., frequency, word length, presence of /s/) influences the perception of a male speaker's sexual orientation. In a series of experiments, we present evidence from English monolinguals, as well as Chinese-English bilinguals, indicative of deeply encoded acoustic and lexical biases, whose manifestation correlates positively with age of acquisition and proficiency in English.

One important finding in current literature is that native speakers of Spanish use gender marking on Spanish articles, such as la and el, to facilitate processing of upcoming Spanish nouns (e.g. Lew-Williams & Fernald, 2010). Conversely, native English speakers (for whom grammatical gender is absent in their first language) who are highly proficient speakers of Spanish and who have demonstrated mastery of the Spanish grammatical gender system, do not behave like native Spanish speakers in this respect. One question, however, is whether this result is modulated by the cognate status of the words. Cognates are words that are similar in form and meaning in the two languages (such as guitar in English and guitarra in Spanish). Here we investigate whether English learners of Spanish can more easily access gender information when words are cognates. We used an experimental eye-tracking technique known as the Visual World Paradigm. In this technique, participants hear instructions to click a picture displayed on a computer screen while their eye movements are being recorded. English-Spanish participants at Penn State were recruited as well as a group of monolingual speakers of Spanish (i.e., the control group) at the University of Granada. Preliminary results show that for the monolingual speakers, cognate status of words does not modulate grammatical gender processing. English-Spanish participants are expected to use grammatical gender only when words are cognates in English and Spanish. This research has important implications because it touches on critical aspects of language learning.

Real-time language processing often requires us to establish grammatical dependencies between units that are grammatically and temporally discontinuous. While we generally achieve this with remarkable speed and accuracy, errors do occur, and these errors can provide rich evidence regarding the structure of the human language processing system. For example, behavioral studies have shown that individuals have difficulty processing subject-verb agreement in both comprehension and production when the subject noun phrase (NP) contains conflicting cues about grammatical number. Specifically, plural nouns embedded in modifying phrases of singular NPs (e.g., 'The key to the wooden cabinets…') lead to an increase in agreement errors in production and a reduction of ungrammaticality effects in comprehension (‘agreement attraction’). Moreover, most studies have found similar profiles of attraction effects in both comprehension and production, leading researchers to propose that the same cognitive mechanisms are responsible for the establishment of agreement dependencies in both tasks (e.g., Nicol, et al, 1997; Severens, et al, 2008; Wagers, et al, 2009). However, here I present evidence from four comprehension experiments investigating agreement attraction in native English speakers which show important asymmetries between comprehension and production. These results indicate that the mechanisms responsible for interference when speaking and reading are at least partly distinct.

Experiment 1 used event-related brain potentials (ERPs) to study agreement attraction. Participants read sentences containing subject NPs with embedded prepositional phrase (PP) modifiers that were either grammatical or ungrammatical (i.e., they showed either correct or incorrect verbal agreement with the singular head noun) and which contained either a singular or plural embedded noun (e.g., 'The key to the wooden cabinet(s) is/*are…'). Results showed that ungrammatical verbs elicited P600 effects, typical of processing grammatical anomalies, but that the P600 was significantly smaller following a plural attractor noun (i.e., an attraction effect). Importantly, there was no difference in brain responses between the two grammatical conditions ('The key to the cabinet(s) is…'). This suggests that, unlike in production, attraction in comprehension differentially affects correct and incorrect outcomes (cf. Staub, 2009).

Experiment 2 expanded on Experiment 1 to investigate structural effect in attraction interference, as behavioral research on language production has shown that the syntactic complexity of the modifying phrase impacts attraction rates. These studies have shown that plural attractor nouns embedded in relative clause (RC) modifiers (e.g., 'The key that opened the cabinets…') lead to significantly fewer errors than those embedded in PPs (i.e., attractor number and modifier structure interact: Bock & Cutting, 1992; Solomon & Pearlmutter, 2004). In Experiment 2, participants read sentences that were created in a 2 (grammaticality: grammatical, ungrammatical) by 2 (attractor number: singular, plural) by 2 (modifier structure: PP, RC) design. Results replicated the ungrammaticality and attraction effects from Experiment 1. There was an additional effect of modifier structure, such that P600s were reliably larger following RCs than PPs. However, this effect did not interact with attractor number, suggesting an asymmetry with production. Experiment 3 followed-up on this result using self-paced reading methodology and a similar experimental design. Results replicated previous self-paced reading studies’ findings of ungrammaticality and attraction effects in ungrammatical sentences, but reading times at the verb were significantly faster following RCs than PPs. Again, this effect did not interact with attractor number. Experiment 4 investigated the relationship of these structural effects with agreement processing by eliminating the need to process verb agreement. Sentences contained an invariant modal verb and were created in a 2 (attractor number) by 2 (modifier structure) design (e.g., The key to the cabinet(s) might be… ). Results were in-line with Experiment 3, such that reading times at the modal were faster following RCs than PPs, suggesting that structural effects are independent of the need to process agreement.

These results show important asymmetries between agreement dependency formation in comprehension and production. Unlike in production, attraction in comprehension differentially impacts grammatical and ungrammatical sentences, and attraction interference in comprehension is not 'clause-bound' as it is in production. Instead, there is a facilitation effect for verb integration following more complex clausal modifiers. These results conflict with complexity-based integration metrics for comprehension (Grodner & Gibson, 2005) and are more in-line with anti-locality effects in verb integration (Vasishth & Lewis, 2006). I argue that the present results are compatible with a content-addressable working memory architecture of sentence comprehension (Lewis et al, 2006).

In this talk I will present an overview of several studies we have carried out in the Baby Research Center in Nijmegen to study the acquisition of various phonological contrasts by Dutch children, using evidence from infant speech perception, word recognition and word production in the second and third year of life.

One important asymmetry that has caused major misunderstandings in the field of phonological acquisition is the gap between children’s knowledge as displayed in perception experiments and the knowledge children bring to the task of language production. For example, while children by the end of their first year of life show knowledge of the sound system of their native language (they seem to know the speech sounds of their language, its phonotactics, stress pattern, etc.), it takes them quite some time before they show that same knowledge in their own productions. Infant speech perception researchers have therefore claimed that perception research provides a better way of tapping children’s grammatical knowledge.

The situation is even more complex: Infants show improved sensitivity to native language contrasts in their first year of life (e.g., Kuhl et al. 2006). However, they show decreased sensitivity to the same contrasts in word-learning experiments in the beginning of the second year of life (e.g., Stager & Werker 1997), although they are still able to discriminate these contrasts in other tasks. This suggests that next to the discrimination of sound contrasts in the pre-lexical stage, in the lexical stage of development another level of perception develops which ignores many phonetic details. We assume that discrimination is based on phonetic properties while word comprehension involves matching those properties to stored phonological representations of words in the mental lexicon. The reduced sensitivity to certain contrasts in word learning might be caused by the nature of early lexical phonological representations.

On the assumption that children use the same lexical phonological representations for word comprehension and word production, we expect to find similar problems in both areas: contrasts that are difficult in comprehension (and hence affect their representation) should also cause problems in production. We show that this is indeed the case. Under such an account there are no major asymmetries between perception and production: both are tightly connected.

In this talk, I describe a body of work exploring translation ambiguity, which occurs when a word in one language has more than one translation into another. For example, the Spanish word "muñeca" translates to both "doll" and "wrist" in English. Our research demonstrates that such ambiguity leads to: (1) slower translation, (2) less accurate translation, and (3) less robust word learning. Furthermore, knowledge that a pair of words share a translation in a later-learned second language impacts the level of perceived relatedness between those words in your first language. For example, native English speakers who learn Spanish as a second language may consider the words "doll" and "wrist" to be more related than native English speakers who do not know Spanish. These findings will be discussed in terms of the ways that the relationship among word meanings across languages influences language learning, processing, and representation.

Classically, patterns of systematic alternations or static distributions in the description of a language have constituted the lion's share of phonological evidence. More recently, laboratory studies have been added to the collection of evidentiary tools for phonological investigation. Both of these approaches have provided the foundations of modern phonological theory, thus their utility is unquestionable. However, they both utilize controlled conditions on data collection which decontextualize language from its natural setting. The goal of this paper is to reaffirm the utility of naturalistic observations of a speech community to phonological investigation, in line with the variationist field of research beginning with Labov (1963).

First, I will briefly outline my assumptions about how observable phonetic variation and change can be related to phonological representations and processes. I will adopt the modular feed-forward approach to phonology (Pierrehumbert, 2006; Bermudez-Otero, 2010), language specific phonetic implementation (Liberman and Pierrehumbert, 1984; Kinston and Diehl, 1994; Boersma & Hamann, 2008), and potentially language specific phonetic alignment constraints (Cohn, 1993; Zsiga, 2000). My model of phonetic change takes place in the language specific implementation of relatively stable phonological objects.

Then, I will walk through an interesting case study, drawn from the Philadelphia Neighborhood Corpus, currently in development (Labov & Rosenfelder, 2011). As of this writing, the corpus consists of 287 sociolinguistic interviews conducted between 1973 and 2010. Speakers in the corpus have dates of birth ranging from 1889 to 1991. The vowel systems of these speakers have been automatically measured (Evanini, 2009, Labov & Rosenfelder 2010), producing 712,822 observations. The Philadelphia Neighborhood Corpus is unique in its size, time depth, focus on a single speech community, emphasis on the sociolinguistic interview, and ethnographic detail.

Specifically, I will focus on a sound change in /ey/. The diachronic raising of /ey/ in non-word-final position was identified early on as a new and vigorous change in Philadelphia, bringing "snake" into close phonetic proximity with "sneak". Utilizing the large volume of observations of /ey/ (45,322), we can investigate this diachronic pattern within many different syllabic, segmental, and morphological contexts. As a result of this detailed diachronic investigation, we can identify and specify an active, synchronic phonological process whereby /ey/ is phonologically peripheralized when followed by a consonant in the same word. This phonological process presumably entered into the dialect during a dialectal realignment of Philadelphia from the South to the North.

Importantly, this phonological process would not have been identifiable without diachronic data. When only looking at speakers grouped together into 10 year date-of-birth bins, no particular pattern of note is evident.

I will conclude with some speculation as to how more social dimensions than geography and diachrony may be leveraged for phonological investigation.

This paper considers a series of far-reaching syntactic changes in the history of Latin, from Proto-Indo-European up to the Romance period. Among others, the following issues are discussed:

1. The shift from SOV to SVO word order
2. The erosion of distinctive nominal inflection
3. The rise of prepositional usage
4. The change in subordination type from non-finite to finite

Despite the broadly structural nature of these changes, we will demonstrate that a pure structurally-based account is inadequate to account for the facts, and that a ranked multi-leveled approach which depends primarily on pragmatic, but also functional, structural and finally typological processes provides a satisfactory set of generalizations. We also discuss the issue of morphological complexity from the perspective of L2 learners, where it has been shown that adult language learners have difficulty with inflection, the very feature which divides fusional from isolating languages.
*Or, “Why the Soviet Union fell

I will share my experiences as a Miccio Award recipient and a PIRE Grad Fellow during Spring '11. As a Miccio Award recipient, I visited the lab of Dr. John Trueswell a the University of Pennsylvania, where I presented my work on the processing of grammatical gender in code-switched speech using an eye-tracking methodology known as the visual world paradigm (Cooper, 1974; Tanenhaus et al., 1995). Beyond sharing the general benefits of visiting another research lab, I will also present new graphs of the eye-tracking data based on new approaches that I learned during my visit. As a PIRE Grad Fellow, I visited the lab of Dr. Teresa Bajo at the University of Granada (Spain). My main focus was to work on an ERP study examining the interaction of verb bias and plausibility in Spanish-English bilinguals. I will explain from the ground up how I compiled the experiment with the help of Dr. Bajo & Dr. Pedro Macizo and present preliminary results from 12 participants.

Traditional models of speech perception suggest that learners of a second language (L2) cannot correctly perceive the differences between native and non-native sounds, which leads to errors in pronunciation (Best, 1994; Flege, 1995). Previous research has also found that motivational factors can affect second language acquisition (Dörnyei, 2005). This paper investigates the abilities of native English speakers in their first three semesters of L2 German to correctly perceive and produce the front-rounded vowels [ø], [œ], [y] and [ʏ]. Perception and production are also discussed in light of individual motivational factors, such as motivation to learn the language, the importance of a native-like accent, and desire to improve accent. Data from an AX perception task, a listen-and-repeat production task, and motivational questionnaire combine to give insights into how these three factors interact. Results indicate that native speakers of American English in the first three semesters of German struggle to correctly produce front-rounded vowels, despite fairly accurate perception of contrasts. They have not yet formed accurate L2 categories in which to place non-native segments, and therefore production of the novel sounds is sporadic and imprecise. Participants’ motivation levels have a significant influence on their perception skills in German. These findings provide a better understanding for German teachers who wish to improve their students’ pronunciation of novel segments.

In an increasingly multilingual world, the study of the cognitive effects of bilingualism has gained much attention in the past few decades. This talk will discuss some of the upsides and downsides of the cognitive consequences of bilingualism by focusing mostly on performance on the Stroop task. I will present behavioural and EEG data showing that bilinguals experience delayed lexical access in their L2 (the downsides), but also enhanced cognitive control (the upsides), and discuss how these processes are modulated by proficiency and language script.

T.B.A.

Past research suggests that word reading skills transfer across writing systems and that bilinguals experience competition between their two print-sound systems. Languages that share a script, such as English and Dutch, differ in their spelling-sound consistency, or the number of pronunciations available for given letter clusters (e.g., consistent: the “-air” in hair, pair; inconsistent: the “-ost” in most, cost). These differences have been shown to affect word-reading strategies. The present research investigated transfer of word-reading strategies in bilinguals who read a second language (i.e., English) that differs in spelling-sound consistency from their first language (i.e., Dutch). Proficient Dutch-English bilinguals made rhyme judgments of visually presented word pairs, while the spelling-sound consistency of the stimuli was varied systematically. Participants were assigned to either a Dutch-English (Experiment 1) or English (Experiment 2) rhyme judgment task. Experiment 1 tested whether word-reading strategies transferred more robustly when both languages were used within a single task than when the task was preformed in the L2 (Experiment 2). My talk will evaluate both behavioral and electrophysiological data to better understand the effects of spelling-sound consistency, orthographic overlap and language context on bilinguals’ granularity (reading unit size) preference.

Despite the fact that Event Related Potentials (ERPs) and functional Magnetic Resonance Imaging (fMRI) allow investigating language processing in real time, they differ in how they do it: ERPs permit to look at language processing with a very high temporal resolution, while fMRI informs on the neural location of language processes. In this talk I will present preliminary data from a series of (old and new) experiments utilizing ERPs and fMRI as convergent measures to investigate language processing in bilinguals. First, I will present (ERPs and fMRI) results on the morpho-syntactic processing of clitic pronouns in native Spanish speakers and English-Spanish bilinguals. Second, I will present data from a set of experiments trying to answer the question of which are the consequences for the L1 when a bilingual speaker is engaged in processing the L2.

A growing body of research has shown that bilingualism enhances aspects of executive control and leads to better performance on a range of cognitive tasks for children and young adults. More recently, this advantage has been shown to extend into older age, demonstrating slower cognitive decline for bilinguals with healthy aging. The present talk will focus on new research that examines changes in the brain that underlie some of these differences. I will report evidence from younger and older adults showing that the regions used by bilinguals to perform certain cognitive tasks are somewhat different than those used by monolinguals, and that the crucial areas in the front part of the brain necessary for performing these tasks are more intact in bilinguals. The presentation will also describe research investigating the memory and cognitive performance of individuals diagnosed with dementia, including Alzheimer's disease, where these protective effects continue to exert an influence on bilinguals.

There is enormous underutilized potential to bring modern research on the behavioral and brain bases of language and cognition to bear on critical issues in education. I will present research concerning the seemingly intractable "achievement gap" in reading between African American children and whites. This gap is not wholly explained by SES or environmental factors, and it increases during the first few years of schooling. One neglected factor is differences in language background. Building on research on first and second language learning and neuroplasticity, we have begun to examine how differences between home and school dialects affect children's classroom experiences. Other factors aside, children who speak a "nonstandard" dialect such as African American English face a more complex learning environment than children who speak the "standard" dialect: they are learning to accommodate the standard dialect while acquiring reading, math and other skills. Because all children are assessed against the same achievement standards, a "gap" results. This research also suggests ways in which the impacts of dialect differences could potentially be ameliorated.

Translation ambiguity has been focus of interest in recent psycholinguistics research. A series of experiments will be presented in which the effect of translation ambiguity was examined using Spanish-Catalan bilinguals. We used a translation recognition task and a translation priming paradigm with lexical decision to look at the effect of several variables: cognate status, dominance of the translation (whether the presented translation is the most frequent one or any other), semantic similarity between the multiple translations (i.e. source of the ambiguity), concreteness, language dominance (Spanish or Catalan) and translation direction (L1-L2 vs. L2-L1). Materials were obtained from an ambiguity database currently in progress. Results showed that translation process was affected by ambiguity so that it was harder to translate words with multiple translations than single-translation ones. In addition, we observed that this effect was found in both cognates and non-cognates although it was larger in the latter. Translation process was hindered when both dominant and subordinate translations were presented, but some differences emerged between them when a translation recognition task was used. In contrast, these differences were not found in a priming experiment indicating that the effect of translation dominance may be due to strategic factors. An exploratory regression analysis on translation recognition data suggests that there was no difference regarding response times between the related (i.e. polysemic) and the unrelated (i.e. homonym) translation pairs. The overall pattern of results indicated that balanced bilinguals such as Spanish-Catalan ones do not show differences in terms of language dominance or translation direction.

The linguistic expression of space draws from and is constrained by basic, probably universal, elements of perceptual/cognitive structure. Nevertheless, there are considerable cross-linguistic differences in how these fundamental space concepts are segmented and packaged into sentences. This cross-linguistic variation has led to the question whether the language one speaks could affect the way one thinks about space – hence whether speakers of different languages differ in the way they see the world. This talk addresses this question through a series of cross-linguistic experiments comparing the linguistic and non-linguistic representation of motion and space in both adults and children. Taken together, the experiments reveal remarkable similarities in the way space is perceived, remembered and categorized despite differences in how spatial scenes are encoded cross-linguistically.

When bilinguals prepare to speak a single word, information about words in both languages is active and appears to compete. The process of initiating speech in each language can be triggered by a range of different events. For example, a bilingual might be asked to name a picture in one language or the other, to speak a word in one language or the other that best fits a definition, or to translate a word in one language into the closest equivalent in the other language. Very little past research has considered how the events that initiate speech planning might differentially reveal this fundamentally competitive process. In the present study, Spanish-English and Chinese-English bilinguals performed a word translation task under conditions in which a word was presented alone and translated into the other language or under conditions in which a picture distractor provided context during translation. Translation performance for the Chinese-English bilinguals was assessed behaviorally and using Event Related Potentials (ERPs). Preliminary results showed that when pictures were present, Chinese-English bilinguals were slower to translate into Chinese, their L1, than into English, their L2. At the same time, ERP data showed that translation from L1 to L2, in the forward direction of translation, was more sensitive to the presence of a semantically related picture than translation from L2 to L1, in the backward direction of translation. I discuss the implications of these results for models of bilingual speech planning and for claims about the ability of proficient bilinguals to exploit semantics directly in the L2.

Some theories of second language (L2) processing (Clahsen & Felser, 2006) claim that late L2 learners never acquire full access to syntactic representations available to native (L1) speakers. Alternative theories account for differences between L2 and L1 processing in terms of reduced availability of cognitive resources in the L2 (McDonald, 2006).

To test these hypotheses we utilized a morpho-syntactic structure that differs between English and Spanish. Spanish clitic pronouns -SpCls- (but not English pronouns) are marked for gender and number and appear before a finite verb. In two experiments, we tested L1 Spanish speakers (Exp1: n=20; Exp2: n=20), and L1 English proficient L2 learners of Spanish (Exp1: n=20; Exp2: n=8).

Experiment 1: The on-line processing of SpCls (in the correct and incorrect position) was tested with a self-paced reading task. Results showed that L1 speakers produced longer RTs at the incorrect clitic site for singular masculine clitics. Other clitics elicited longer RTs on the following word. L2 learners showed an effect at the clitic site but no spillover effect, suggesting that L2 speakers are able to access the morpho-syntactic representation of clitics and exploit it to resolve ungrammaticality. However, cognitive constraints may limit their ability to use this information to make predictions about upcoming information.

Experiment 2: We used Event-Related Potentials to examine the real-time course of SpCl processing. Participants read sentences in which clitics varied in correctness for gender, number, or both. Initial results revealed that both L1 and L2 showed a larger positivity in the 500-700 ms window for the incorrect clitics suggesting sensitivity to the online processing of morphosyntactic information.

Taken together, these results suggest that there are not hard constraints preventing late bilinguals from accessing L2 grammatical information, but rather cognitive consequences of processing the non-native language distinguish the L2 users from native speakers.

The village of San Basilio de Palenque, in the department of Bolivar, Colombia, is a village of approximately 5,000 people founded in the early 17th century by runaway slaves, cimarrones, escaped from the Spanish in the port city of Cartagena. In this initial phase, the cimarrones developed a creole language, called lengua palenquera (LP) based on the Spanish lexicon with phonology, a pronoun system, and a verbal system which are heavily influenced by the various languages these former slaves had carried with them from Africa.

The community has been bilingual from the beginning, always maintaining Spanish alongside LP for purposes of working in and interacting with other Colombian communities, and in more recent generations LP was almost lost due to social pressures from others who viewed LP as little more than poorly spoken Spanish. In the 1980s, an effort began to revive LP by teaching it in San Basilio’s public schools. The result is a present-day population of older lengua-Spanish and younger Spanish-lengua bilinguals, all of whom engage in frequent and seamless code-switching in their everyday interactions with one another and with language researchers arriving from abroad.

The present paper investigates the pattern of code-switching observed in San Basilio de Palenque, Colombia using the diagnostic features identified by Muysken (2000) and data collected by the researcher and colleagues during a trip to Colombia in July 2011. Using this data (and lack thereof, in some cases), it is observed that alternation between LP and Spanish is not as rampant as it has been treated in past work. Following the argument in Schwegler (1996), those alternations which do occur are seen as code-switches realized by highly proficient speakers of both languages. Where code-switching occurs, it follows Muyksen’s (2000) pattern of congruent lexicalization, where lexical items from both languages are inserted into a shared grammatical structure. It is also proposed that there are different processes at work across different generations of Palenqueros.

Additionally, the present data sheds new light on the renewed strength of lengua within the community and the success of the ethno-education program which has been developing over the past twenty years.

The present paper examines a number of issues concerning clause-final negation structure in Palenquero Creole, a Spanish-based creole language spoken in San Basilio de Palenque, Colombia. Although Palenquero derives the great majority of its lexical items from Spanish, its morpho-syntax differs in important ways from its lexifier language, perhaps most notably in terms of negation structure. That is, while in Spanish the negation marker no appears pre-verbally, as in “no voy a hablar más”, the preferred structure in Palenquero Creole places the negator nu post-verbally and more surprisingly, clause-finally, as in “í bae kondbesá má nu” (trans. for both: ‘I am not going to talk more’). Typologically, clause-final negation is quite marked, in that it occurs infrequently across different languages. From a functionalist perspective, this suggests that the structure is disfavored by pressures of language usage and processing (e.g. Bates & MacWhinney, 1989; Givón, 1991). In other words, late negation may be difficult to process. I aim to explore the extent to which the potential difficulty of late negation might be mitigated by prosodic cues, as compared to the realization of prosody in other negation structures and in non-negated utterances. As I continue this research in the future, I am also interested in determining to what extent the semantic (un)expectedness of the upcoming negation might influence prosody. Analysis for the current study was conducted using data from sociolinguistic interviews collected in Colombia in July 2010. Although the preliminary results are not conclusive, similar work on negation in Brazilian Portuguese (Schwenter, 2005; Armstrong & Schwenter, 2008) suggests that such an interaction between prosody and syntactic structure is indeed a plausible one, and that further examination of this and other data may reveal differential prosody in late negated clauses.

A bilingual's cognitive architecture is highly interactive and dynamic, both within and across languages. In this talk, I will show that knowing two languages changes spoken language comprehension and yields co-activation of lexical items across both languages. Using eye-tracking and mouse-tracking data, I will suggest that bilinguals effectively recruit both bottom-up and top-down mechanisms to efficiently and seamlessly integrate information across modalities when resolving ambiguity during comprehension. Bilinguals' domain-specific experience with cross-linguistic competition shows a relationship to domain-general executive function, suggesting that bilinguals may be particularly adept at inhibiting irrelevant information. One consequence of this greater inhibitory experience is a bilingual advantage in novel language learning -- compared to monolinguals, bilinguals are better at learning a new language and show less competition from the native language when using a newly-learned language. These differences in language processing, language learning, and inhibitory control suggest fundamental changes to linguistic and cognitive function as a result of bilingualism.

Being able to process language quickly is a vital skill we rely on for human interactions. Electroencephalography (EEG) research indicates that lexico-semantic information is accessed within 200 ms (e.g. Dell'Acqua et al., 2007). Lower proficiency in a second language (L2) relative to a first (L1) leads to delays in lexical processing speed in L2, as proposed by the temporal delay assumption of the BIA+ model (Dijkstra & van Heuven, 2002; van Heuven & Dijkstra, 2010). The reduced frequency hypothesis (Pyers et al., 2009) proposes that the L1 is also delayed due to lower frequency of use relative to monolinguals The current study investigates these hypotheses in the context of automatic reading by directly comparing English monolinguals and Chinese-English bilinguals' L1 and L2 on a Stroop task. Stimulus onset asynchrony (SOA) was manipulated to provide further information on the automaticity of word reading. Word stimuli elicited an N170 component, which is related to orthographic processing (Bentin et al., 1999), in all SOAs, indicating automatic lexical processing in all groups. At the N170 peak, differences emerged between words and control stimuli in monolinguals and bilinguals' L1, signifying that early lexical processing occurred at the same latency for the native languages despite large orthographic differences. In bilinguals' L2, however, word and control waveforms diverged significantly later, indicating delayed lexical processing. These results provide neurophysiological evidence for the temporal delay assumption but not the reduced frequency hypothesis, and confirm that early lexical processing is automatically activated but significantly delayed in the second language.

Most research on language acquisition has assumed an idealized input which lacks any variability, other than the overall variety of sentences allowed by a particular grammar. However, it is well known that this idealization is just that, and that real language data is variable. The linguistic data that each child is exposed to is subject to all types of linguistic and extra-linguistic (dialect, gender, age, speech style) variation, which is not always categorical in nature. If we take this fact into consideration, the acquisition problem becomes much harder but also more realistic. Variability (within and across speakers), even though probabilistically constrained, can add more ambiguity to the input the child is exposed to, making some input unreliable for a particular grammatical generalization although perfectly reliable for learning some other property of the language. In this presentation I will present research that examines the effect of input variability on acquisition by comparing two highly similar dialects that differ minimally in terms of the frequency and reliability of a particular feature. By comparing the inputs and the rates of acquisition of this feature in the two varieties, we can begin to understand the complex relationship between input data and grammar acquisition.

The right hemisphere's role in processing metaphors has been debated. While clinical research suggests that damage to the right hemisphere impairs figurative language comprehension, results from imaging studies have been mixed. Additionally, the Graded Salience Hypothesis proposes that other factors such as novelty and context, rather than figurativeness per se, may influence hemispheric recruitment. In two separate fMRI studies, we examined how novelty and context influenced hemispheric recruitment in processing figurative and literal sentences. In experiment 1, all metaphors and novel stimuli elicited activation in bilateral frontal and left temporal regions. Additionally, all metaphors engaged the right temporal pole. In experiment 2, a main effect of figurativeness (metaphors > literal) was found in left hemisphere regions only. A main effect of congruence (congruent > incongruent) revealed activations in bilateral frontal and temporal regions, and left dorsal medial prefrontal cortex (DMPFC). In our first experiment, although an influence of novelty was found, the right hemisphere's sensitivity to familiar metaphors suggests that even relatively familiar metaphors still require additional semantic integration. Our second study demonstrated that in the presence of additional context, metaphors and literal sentences did not differentially engage the right hemisphere. In contrast, processing coherent discourse compared to incoherent discourse, regardless of the figurative or literal aspect of the text, engaged right inferior frontal and temporal gyri, and left hemisphere regions including DMPFC. These results are partially consistent with the Graded Salience Hypothesis, highlight the strong influence of novelty and context on language, and suggest that in a wider discourse context, congruence has a much stronger role in right hemisphere recruitment than figurativeness.

A finding in recent studies of bilingual word recognition is that it is impossible to restrict activation to one language alone, even when reading in sentence context. The activation of the language not in use persists under almost all conditions except when sentences are highly constrained semantically. Here we asked whether a similar effect of constraint would be observed when sentences are syntactically specific to one of the bilingual?s two languages. Proficient Spanish-English bilinguals read sentences in each language that contained a to-be-named cognate or matched control word. Half of the Spanish sentences contained syntax that was structurally specific. English sentences were translations of the Spanish sentences but were not syntactically unique to English. Results indicate that the cognate effect in word naming was not always reduced following syntactically specific sentence context in Spanish. The utilization of syntactic constraint appears to depend on language dominance. These findings have implications for claims about bilingual word recognition, language immersion, and language dominance.

Word learning is complicated. Learners face high mapping ambiguity in cluttered visual and acoustic environments, and research with monolinguals has demonstrated a primacy effect in the learning of word-object mappings. Specifically, mutually exclusive mappings are preferred even in contexts where many-to-one mappings are available. The situation may be even more complicated for bilinguals, who must learn that two words, one from each of their languages, map to the same referent (e.g, "dog" and "chien" both refer to a four-legged furry creature). These observations present three clear questions for research in word learning: 1) how can we describe the learning mechanism necessary for overcoming the ambiguity encountered in typical learning environments; 2) how does a learner override the preference for mutual exclusivity in word learning to establish many-to-one word-object mappings, and 3) how do we relate the work on monolinguals to our understanding of word learning in bilinguals? The research presented here will address these questions through the lens of statistical learning.

Previous research has demonstrated that monolingual learners are sensitive to the distributional properties of linguistic input. Specifically, learners are able to detect structural boundaries (i.e., transitions between segments, syllables and words) and transitions between languages by tracking contextual and phonological cues embedded in acoustic input. Yu and Smith 2007 developed a cross-situational learning paradigm in which the possible referents of words are established via the statistical properties of their distribution in the input. They demonstrated that learners are able to track the co-occurrence statistics of words and possible referents across contexts, and thus successfully link words and referents. In the present research, we propose to investigate how the addition of contextual and phonological cues to the acoustic input offered in the cross-situational paradigm modulates learners' established preference for mutually exclusive one-to-one mappings of words and objects. If the cross-situational learning mechanism is also sensitive to cues identifying the particular language set from which co-occurrence statistics are drawn, we hypothesize that such cues should allow learners to override mutual exclusivity and learn the many-to-one mappings evident in the linguistic knowledge of all bilinguals. To date, we have replicated the findings that 1) the cross-situational learning paradigm is able to support the learning of novel word-object mappings and 2) mutual exclusivity is an operant constraint on word-learning in bilingual contexts. We have begun a series of experiments investigating the strength of various contextual and phonological cues in overriding this constraint.

Patterns of priming for regularly inflected and form similar prime-target pairs provide a window on morphological processing in English as a first and as a second language. I will present results from L1 speakers of Dutch, Serbian and Chinese as they perform a cross modal lexical decision task in English. These conditions permit a strict test of the claim that, because they lack the grammar (procedural knowledge) to analyze morphologically complex word forms, L2 users rely on lexical (declarative) knowledge to recognize morphologically complex words. Further, the inclusion of three L1s with different morphological and phonological structures presents the opportunity to detect different patterns of interaction between L1 and L2. Across levels of L2 proficiency, we ask whether command of inflectional morphology as revealed by magnitudes of morphological facilitation in L2 differ as a function of L1. Consistent with the claim that particular dimensions of similarity between various L1 and L2 (poverty of morphological complexity in Chinese compared to Dutch or Serbian) can affect processing in L2, cross modal findings show variation in the magnitude of facilitation across L1 speakers of Chinese and Serbian when proficiency is controlled. Also novel is a finding that Dutch L1-English L2 speakers appear to show greater morphological facilitation relative to a form control when English primes are pronounced in an American than in a Dutch accent; our generally less proficient Chinese participants did not show this pattern. Similarities and differences between patterns of morphological facilitation in L1 and L2 are discussed.

A number of studies have suggested that argument phrases in sentences are easier to process than non-arguments. The processing advantage of arguments is thought to arise from their lexical specification. Non-arguments, or adjunct phrases, in contrast, may depend on a syntactic linkage to the phrase they modify. This talk will explore whether preliminary data from a self-paced listening task support these hypotheses. The differentiation of argument and adjunct processing will also be considered in light of the procedural deficit hypothesis (Ullman & Pierpont, 2005), which suggests that specific language impairment stems from a weakness in the procedural memory system.

This presentation focuses on Media Lengua de Imbabura (MLI), which is a mixed language spoken in several communities located in the Northern part of the Ecuadorian highlands. I will explore and describe the social factors that could have possibly leaded to the emergence of MLI, and it will also address the linguistic question of how MLI can inform our view on the classification of adpositions. If the definition of mixed language is based on a clear division between lexical and grammatical items, then looking at how adpositions are treated in MLI can give us some insight about the character of this part of speech. I intend to present a functional analysis of adpositions that would try to answer the following questions:

1. How are adpositions treated in MLI?

2. What can this treatment tell us about the general classification of adpositions?

In order to answer these questions I will be using a functionalist approach.

As in other locales in the United States, Spanish in New York has been studied for the most part in order to highlight the English-origin features found in it as a result of the forces of language contact. While there is no doubt that contact plays a real role in shaping Spanish in New York, the attention paid to this element has obscured two other important contributors to the form of Spanish in the City, namely (a) dialectal leveling and (b) structural continuity with Latin American Spanish. Almost as much as does the presence of English, it is the presence of ways of speaking Spanish other than one’s own that is giving form to Spanish in the City. Yet these two elements, language contact and dialectal leveling, are not enough. For these two considerations are about how Spanish in New York differs from the Latin American reference lects. But equally important are the ways in which Spanish in the City remains the same as the Spanish of Puerto Rico, Mexico, and the other places whose speech ways have been imported into the five boroughs of the Big Apple. In order to establish with some precision the boundaries of contact, leveling, and continuity, a micro analysis is needed of a single, variable linguistic feature, analyzed at the correct level of abstraction. The feature studied in the present project is the variable use of subject personal pronouns, that is, the alternation between presence and absence of the pronoun, as in canto ~ yo canto ‘I sing’, cantas ~ tú cantas, ‘you sing’, etc.. The project relies on the methods and theoretical machinery of variationist sociolinguistics, using variable hierarchies and constraint hierarchies to bring out the coexistence of contact, leveling and continuity as shapers of Spanish in the City. The project pays special attention to the Spanish of New York-raised, second-generation Latinos, and offers a commentary on the widely accepted notion of incomplete acquisition as a characterization of their linguistic competence.

The 2010 recipents of the Miccio Travel Award will describe their experiences and give tips on how to apply and how to get the most from the opportunity.

Givenness and pronominality are highly correlated in adult speech, and they tend to align in adult dative constructions such that THEME-first constructions (e.g., Give it to the man) are more likely when the THEME is given and pronominal, and RECIPIENT-first constructions (e.g., Give him the hat) are more likely when the RECIPIENT is given and pronominal (Bresnan & Nikitina 2009). While previous work has shown that givenness and pronominality are also correlated in child speech (e.g., Matthews et al. 2006), there has been little research considering how these two factors influence early syntax. The current studies reveal that both givenness and pronominality play a role in the dative construction choices of young English speakers. In two studies, I prompted four-year-olds to describe video clips in different discourse conditions: one condition with a given THEME, one with a given RECIPIENT, and a control condition where neither argument was given. Discourse condition strongly influenced word order. THEMES were more likely to be first in the THEME-given condition. This effect was categorical in Study 1 and highly robust in Study 2 (p < .0001). Similarly, RECIPIENTS were more likely to be first in the RECIPIENT-given condition, though less so (Study 1: p < .05; Study 2: ns). The differences between THEMES and RECIPIENTS and between Studies 1 and 2 are largely attributable to pronominality. Given arguments were generally pronominalized, and pronouns were generally ordered first. Importantly, children always ordered THEMES-pronouns first (Give it to the man), while they sometimes ordered RECIPIENT-pronouns last (Give the hat to him), hence the stronger effect of THEME-givenness. And given information was pronominalized more in Study 1 than Study 2, likely because Study 1 participants heard the given information mentioned and saw a picture of it, while Study 2 participants only heard the information. These studies highlight the continuity between child and adult language production and underline the importance of incorporating information about discourse status and the type of referring expressions into models of early syntactic development.

I will present a collection of our recent studies investigating cognitive underpinnings of poor language learning in children and adolescents. The first set of studies investigated the component skills involved in completing verbal working memory tasks. Many current theories argue that limitations in working memory capacity result in incomplete language learning. However, performance on the tasks used to measure working memory appear to involve several component skills that go beyond maintaining verbal computations in working memory. Our studies identified some of these component skills, and therefore, cast a doubt on usefulness of the construct of working memory in explaining individual differences in language development. I will also discuss emerging new work investigating the role of implicit learning mechanisms in explaining poor language development. I will finish with my vision of where this pursuit should be heading towards.

Until recently, research on language and its cognitive interface focused almost exclusively on monolingual speakers of a single language. In the past decade, the recognition that more of the world's speakers are bilingual than monolingual has led to a dramatic increase in research that assumes bilingualism as the norm rather than the exception. This new research investigates the way in which bilinguals negotiate the presence of two languages in a single mind and brain. A critical insight is that bilingualism provides a tool for examining aspects of the cognitive architecture that are otherwise obscured by the skill associated with native language performance. In this talk, I illustrate the ways in which bilingualism reveals the architecture and mechanisms for language processing and their neural basis.

Although the adult brain is far from being fixed, the types of experience that promote learning and brain plasticity in adulthood are still poorly understood. Surprisingly, the very act of playing action video games appears to lead to widespread enhancements in visual skills in young adults. Action video game players have been shown to outperform their non-action-game playing peers on a variety of sensory and attentional tasks. They search for a target in a cluttered environment more efficiently, are able to track more objects at once and process rapidly fleeting images more accurately. This performance difference has also been noted in choice reaction time tasks with video game players manifesting a large decrease in reaction time as compared to their non-action-game playing peers. A common mechanism may be at the source of this wide range of skill improvement. In particular, improvement in performance following action video game play can be captured by more efficient integration of sensory information, or in other words, a more faithful Bayesian inference step, suggesting that action gamers may have learned to learn.

A bilingual's cognitive architecture is highly interactive and dynamic, both within and across languages. In this talk, I will show that knowing two languages changes spoken language comprehension and yields co-activation of lexical items across both languages. Using eye-tracking and mouse-tracking data, I will suggest that bilinguals effectively recruit both bottom-up and top-down mechanisms to efficiently and seamlessly integrate information across modalities when resolving ambiguity during comprehension. Bilinguals' domain-specific experience with cross-linguistic competition shows a relationship to domain-general executive function, suggesting that bilinguals may be particularly adept at inhibiting irrelevant information. One consequence of this greater inhibitory experience is a bilingual advantage in novel language learning -- compared to monolinguals, bilinguals are better at learning a new language and show less competition from the native language when using a newly-learned language. These differences in language processing, language learning, and inhibitory control suggest fundamental changes to linguistic and cognitive function as a result of bilingualism.

I will describe a pilot project that is designed to compare the usefulness of processing-based vs. knowledge-based language measures for predicting theory of mind ability in preschoolers. I will outline some of the theoretical issues concerning relationships between language and social cognition. The main aim of the presentation is to provoke discussion about -the theoretical and practical issues involved in the pilot study and a proposed “real” study.

Most research on language acquisition has assumed an idealized input which lacks any variability, other than the overall variety of sentences allowed by a particular grammar. However, it is well known that this idealization is just that, and that real language data is variable. The linguistic data that each child is exposed to is subject to all types of linguistic and extra-linguistic (dialect, gender, age, speech style) variation, which is not always categorical in nature. If we take this fact into consideration, the acquisition problem becomes much harder but also more realistic. Variability (within and across speakers), even though probabilistically constrained, can add more ambiguity to the input the child is exposed to, making some input unreliable for a particular grammatical generalization although perfectly reliable for learning some other property of the language. In this presentation I will present research that examines the effect of input variability on acquisition by comparing two highly similar dialects that differ minimally in terms of the frequency and reliability of a particular feature. By comparing the inputs and the rates of acquisition of this feature in the two varieties, we can begin to understand the complex relationship between input data and grammar acquisition.

After an extended period of time in a second language (L2) environment, some bilinguals may become more proficient in the L2 than in the native language. This switch of language dominance can be observed under a variety of circumstances and at a range of different points in the lifespan, e.g., following immigration or after growing up in a household that speaks a minority language and then entering school in which instruction is delivered in the majority language. Although there has been some past research on immersion experience, there has been little attention paid to switches of language dominance, either for language processing or for their cognitive consequences. The present study examined language processing in native Spanish speakers living in the US who have become proficient in English as the L2. Immersed in a largely monolingual environment, some bilinguals have become dominant in English, thereby switching language dominance. Spanish-English bilinguals who have switched language dominance in this context were compared to Spanish-English bilinguals who maintained dominance in Spanish and also to native English speakers with Spanish as the L2, and to monolingual speakers of English. Participants were tested on a set of language processing tasks and also a set of cognitive measures. The language processing tasks included picture naming and verbal fluency, to assess performance at the lexical level, and a sentence processing task, to assess attachment preferences. The cognitive tasks included measures of working memory and inhibitory control. Preliminary results suggest a dissociation between the manifestation of language dominance at the lexical and sentential levels. Spanish-English bilinguals who have become English dominant for lexical-level tasks may retain some Spanish-specific preferences in sentence processing. Furthermore, there appear to be separable influences of switching language dominance and language immersion per se. The results of this study have theoretical implications for claims about the plasticity of the language system across an individual’s life experience. They also introduce a set of questions concerning the way in which past research has categorized bilinguals on the basis of native language status alone.

In several regions of South America, Spanish is in contact with the Native American language Quechua, and beginning with the Spanish colonization in the 16th century a variety of stable as well as transitory interlanguage varieties have developed, in highland regions of Ecuador, Peru, Bolivia, and southern Colombia. Quechua (known as Quichua in Ecuador) is characterized by a three-vowel system, traditionally represented as /ɪ/-/a/-/ʊ/, and when Quichua speakers attempt to acquire the Spanish five-vowel system (/i/-/e/-/a/-/o/-/u/), the mid-high vowel distinctions (/i/-/e/ and /u/-/o/) are seldom completely mastered. Popular opinion holds that Quichua-dominant speakers actually interchange high and mid vowels, but in reality the situation is much more complex. Previous research has examined only elicited individual words in laboratory settings. The present research project involves spontaneous speech collected among elderly Quichua-Spanish bilinguals in northern Ecuador who acquired Spanish in late adolescence or early adulthood, who have received no formal education in any language, and who continue to speak more Quichua than Spanish on a daily basis. An examination of the vowel spaces of Quichua-influenced Spanish interlanguage reveals no simple transference but rather a complex array of expanded vowel spaces that correspond neither to Quichua nor to Spanish. In the expanded and still relatively amorphous vowel spaces corresponding to the Spanish /i/-/e/ and /u/-/o/ oppositions, preliminary results suggest emergent processes that have the cumulative effect of reducing the “entropy” of the vowel dispersions, but which also contribute to the popular perception of “mix and match” vowel confusion. The Quichua-Spanish vowel data are evaluated in the light of several models of vowel production/perception and second-language acquisition of phonological contrasts.

Learning grammatical gender is one of the most difficult aspects of learning German as a second language. Many speakers continue to have difficulties in gender assignment and use even when they are overall very proficient in German. This study looks at three aspects of gender acquisition at very early stages of learning, previous foreign language experience, sensitivity to phonological regularities, and patterns in the classroom input. I will present data from a grammatical gender assignment task performed by second semester students of German, and an analysis of the patterns of gender in an introductory text book. Results show that all participants are sensitive to some phonological regularities, regardless of previous foreign language experience.

Research in our laboratory focuses on how the mental representation of L1 and L2 develops and how multiple linguistic systems compete as a function of age of acquisition and language proficiency. This talk will present a series of ongoing behavioral and fMRI studies that investigate (1) whether bilingual lexical activation is modulated by external contextual cues such as facial features of the interlocutor, (2) whether mathematical processing in L1 versus L2 involves distinct neural substrates as a result of difficulty of task and L2 proficiency, (3) whether working memory and executive control can distinguish faster L2 learners from poor L2 learners in a set of novel word and grammatical learning tasks, and (4) whether object naming in L1 shows negative transfer after learning of L2 as a function of the congruency of the naming patterns between the two languages. Implications of these studies are discussed in light of current theories of bilingualism and second language acquisition.

One of the long standing controversies in generative and experimental treatments of predicates (regardless of the theoretical approach) is the debate centering on the internal wealth of detailed information or lack thereof that predicates contain. To put it simply, do predicates - called √roots by Pesetsky (1995) - determine the argument structure of the syntax (i.e., a projectionist view) or, in contrast, is it the structure of the syntax that is responsible for endowing an impoverished √root with semantic information based solely on its position in the structure (i.e., a constructionist view)? In this presentation, I take a closer look at über-prefixing in German (and related languages). Following Risch (1995), McIntyre (2003), and Putnam (to appear), I demonstrate that scalar implicatures are lexicalized in √roots on a case by case basis (based largely on the aktionsart-type of the predicate) (see Rappaport-Havov 2009 for a similar proposal). Finally, I discuss the morpho-syntactic consequences of scalar properties in connection with these über-verbs in German and beyond. For example, consider the following examples from German:

(1) Sergej überißt *(sich) an Bananen.

Sergej over-eats Refl on Bananen_DAT

‘Sergej overeats on bananas/Sergej eats too many bananas.’

(2) Der Hund überbellt *sich/die Katze.

the dog over-barks REFL the cat

‘The dog outbarks the cat.’

Although both über-verbs involve a scalar event, example (1) requires the overt presence of a weak pronominal reflexive, whereas the second example (2) does not. The situation in German for example (1) contrasts with data from English – see (3) below – where the presence of a reflexive in a similar scalar context is ungrammatical:

(3) Richard overeats (*himself) (on pizza).

Contrary to most treatments of these verbs, I argue that the reflexive in these constructions does receive a distinct theta-role interpretation (i.e., Standard). Furthermore, I suggest that the presence of the weak reflexive in German and other related languages is the lexicalization of a pro-argument in a degree phrase (= DegP) situated within the verb phrase.

In this presentation, I introduce a new perspective on questions important in the field of bilingualism by exploiting the exquisite temporal resolution of Event-Related Potentials (ERPs). First, I demonstrate how late Chinese-English bilinguals unconsciously translate English words into their Chinese native equivalents, whether English words are presented auditorily or visually. Second, I provide evidence is support of the linguistic relativity hypothesis by showing an effect of language-specific colour terminology on colour discrimination in Greek-English bilinguals. Third, I show how millisecond-by-millisecond tracking of brain activity in relation to behavioural output enables us to establish the time of lexical access in picture naming by studying the cognate and lexical frequency effects in Spanish-English bilinguals and the Semantic Competitor Inhibition Effect (Howard et al. 2006) in monolinguals. In conclusion, I discuss how ERPs allow us to track crucial stages of mental processing well before a behavioural response is observed.

Written language is an extraordinary human invention that has allowed for the communication and accumulation of knowledge across time and geography and, in so doing, has revolutionized human history. However, in evolutionary terms it has entered the human repertoire only very recently, without the opportunity to carve out its own territory within the human genome. This raises a number of questions, including: How has the brain accommodated written language processes and representations? How are these related to those of evolutionarily older skills such as spoken language, object recognition, working memory? What is that we know when we know how to read and write words? In this talk I will review findings from behavioral research involving cognitive neuropsychological studies of individuals with acquired written language deficits, as well as fMRI research with neurologically intact individuals. I will argue that both approaches make use of similar experimental logic, including such things as: dissociation, association, facilitation, parametric variation, analysis of similarity. These lines of research are beginning to converge in uncovering the functional architecture of the written language processing system and how it is related to other cognitive systems. Furthermore, both lines of research are increasingly able to reveal the richness and internal complexity of orthographic representations.

Bimodal bilinguals, fluent in a signed and a spoken language, exhibit a unique form of bilingualism because their two languages access distinct sensory-motor systems for comprehension and production. When a bilingual’s languages are both spoken, the two languages compete for articulation (only one language can be spoken at a time), and both languages are perceived by the same perceptual system: audition. Differences between unimodal and bimodal bilinguals have implications for how the brain might be organized to control, process, and represent two languages. In this talk, I highlight recent results that illustrate what bimodal bilinguals can tell us about language processing and about the functional neural organization for language.

An extensive body of research claims that labels facilitate categorisation, highlight the commonalities between objects and act as invitations to form categories for young infants before their first birthday. While this may indeed be a reasonable claim, we argue that it is not justified by the experiments described in the research. We report on a series of experiments that demonstrate that labels can play a causal role in category formation during infancy. Ten-month-old infants were taught to group computer-displayed, novel cartoon drawings into two categories under tightly controlled experimental conditions. These findings demonstrate that even before infants start to produce their first words, the labels they hear can override the manner in which they categorise objects. Yet little is known regarding the nature of the mechanisms by which this effect is achieved. We further describe a neuro-computational model of infant visual categorisation, based on self-organising maps, that implements the unsupervised feature-based approach. The model successfully reproduces experiments demonstrating the impact of labelling on infant visual categorization reported in Plunkett et al. (2008). The results suggest that early in development, say before 12-months-old, labels need not act as invitations to form categories nor highlight the commonalities between objects, but may play a more mundane but nevertheless powerful role as additional features that are processed in the same fashion as other features that characterise objects and object categories.

Language-internal and –external processes in the formation of spatial prepositions in Papiamentu Papiamentu, a Romance-based creole, has a rich, established prepositional system in contrast to other creole languages (Kouwenburg & Murray, 1994). The great majority of these prepositions appear to be transparently derived from their Romance counterparts. However, I will examine two spatial prepositions—riba (>Sp., Port. arriba) and for di (>Port. fora de, Sp. fuera de)—which have semantically expanded to take on additional meanings not exhibited by their Romance counterparts. I will argue that these prepositions exhibit two different processes by showing language-internal processes at work in the expansion of riba and reviewing substratum influence (i.e. language-external) in the case of for di (Maurer, 2005). Finally, I highlight the need to examine lexemes individually as they ostensibly follow similar paths of grammaticalization. Creole languages in general offer a clear warning of attributing synchronic outcomes to one catch-all mechanism.

Bilingual aphasia, defined as a loss of one or both languages in bilingual individuals that results from left hemisphere damage, is of increasing interest worldwide because half the world’s population is bilingual. In the United States, the elderly Hispanic population is the fastest growing ethnic minority (Bureau of the Census, 2006). However, current research on bilingual aphasia cannot inform or recommend the optimal rehabilitation for bilingual aphasic patients (Roberts & Kiran, 2007). For instance, it is not known whether or not rehabilitating one of the patient's languages is sufficient, nor to what extent cross-language transfer occurs after rehabilitation. Several factors contribute to the paucity of research in this area: the multitude of possible language combinations in a bilingual individual, the relative age of acquisition (AoA) and proficiency of the two languages of the bilingual individual, and the effect of focal brain damage on bilingual language representation. In this talk, I will focus on three broad issues, 1) what we understand about brain representation of two languages in normal and brain damaged bilingual individuals, 2) what we understand about the cognitive control of lexical access in bilingual aphasia through analysis of cross-language errors and 3) what we know about cross-language transfer subsequent to rehabilitation in one language. Using four experimental methodologies, fMRI, computational modeling, behavioral analysis of language production and single subject treatment designs, I will provide some insight into the complexities of bilingual aphasia rehabilitation and the various factors that contribute to cross-language transfer in these patients.

What factors affect the coding of two languages in one brain? For over 100 years, researchers have suggested that age of acquisition (when) vs. proficiency (how well) in a particular language play a role in its neural representation. Recent work in my laboratory has explored the influence of these two variables in bilingual language processing using fMRI. Studies have also extended this work by looking at these two factors in monolinguals and in motor skill processing in athletes. The similarities across these domains provide compelling evidence of the link between language and motor skill learning. They are also consistent with an emergentist view in which neural representations arise from a series of interactions at multiple levels. The implications of this conceptualization of language for clinicians and educators alike will be discussed.

What factors affect the coding of two languages in one brain? For over 100 years, researchers have suggested that age of acquisition (when) vs. proficiency (how well) in a particular language play a role in its neural representation. Recent work in my laboratory has explored the influence of these two variables in bilingual language processing using fMRI. Studies have also extended this work by looking at these two factors in monolinguals and in motor skill processing in athletes. The similarities across these domains provide compelling evidence of the link between language and motor skill learning. They are also consistent with an emergentist view in which neural representations arise from a series of interactions at multiple levels. The implications of this conceptualization of language for clinicians and educators alike will be discussed.

Theories of language production typically assume that at all levels of processing non-target representations associated with the target are partially activated. For example, at the lexical level, semantic associates are typically assumed to be activated (for target CAT, words like RAT, DOG, etc.) To cope with potential interference from these representations, theories typically incorporate selection mechanisms that serve to enhance target processing (e.g., boosting the activation of a node representing CAT). Over the past two decades, an extensive body of work (in both mono- and multilingual production) has shown that at the lexical level selection is not discrete. Although selection processes extensively enhance target activation, non-target representations (both within and across languages) remain partially active, influencing subsequent phonological processing (e.g., mixed semantic-phonological neighbors such as RAT facilitate phonological retrieval for target CAT).

In this talk, I'll review recent evidence that non-discreteness extends to phonological and phonetic processing. In monolinguals, gradient variation in the activation of phonological representations influences the phonetic realization of targets. In multilingual production, interaction between the speaker's sound systems at the phonetic level is modulated by gradient variation in the activation of phonological representations. I'll discuss the implications of these findings for phonological and phonetic processing in both mono- and multilingual production.

Despite a variety of logical and empirical problems, many researchers believe that language learning is limited by a critical period. The unified version of the Competition Model presents a way of accounting for age-related differences in language learning abilities that does not rely on critical periods, but instead on first language entrenchment, competition between multiple languages, and changing patterns of social integration into a new language community. The analysis has led to a variety of experiments designed to evaluate ways of improving L2 learning in adulthood.

Studies of both first and second language acquisition have largely focused on the acquisition of form over meaning. While comprehension studies indicate that language learners' understanding is not always adult- or target-like, surprisingly little is known about the nature of the differences, the details of children's and adult L2 learners' semantic systems once forms are in use, and when and what changes take place. In this talk I will present three studies exploring what child and adult language learners' gestures reveal about their verb meanings. The target domain is of that of placement (e.g., putting a cup on a table), which is lexicalized differently crosslinguistically. The first study shows how differences in placement verb meanings in Dutch and French are reflected in two distinct patterns of adult gesture use. The second study examines Dutch four- to five-year-old children's acquisition of placement verbs demonstrating that their placement gestures change systematically as their placement verb meanings develop. The last study illustrates different gesture patterns in adult Dutch learners of L2 French depending on influences of the L1 and different degrees of semantic reorganization. Together the studies support the notion that speech and gesture form an integrated system as revealed (a) in robust crosslinguistic differences in gestural practices parallel to differences in speech, and (b) in similar parallel differences across modalities in development. The integrated nature of the systems further means that gestures open a new window on details of semantic representations; and that they can shed light on the process of acquisition by revealing shifts in such representations.

Research on code-switching has progressed to the extent that there are now several competing models attempting to account for the patterns found in conversational data from bilinguals. One of the goals of our research programme at Bangor is to critically evaluate these competing models rather than to work within only one theoretical framework. The purpose of this talk is to defend the goal of critical evaluation in the face of the argument that two theories are never comparable, or what philosophers of science have called ‘incommensurability’. I seek to show in particular that the critical evaluation of two theories which at first sight appear not to be conducive to comparison can lead to new insights, including the redefinition of concepts and the generation of new hypotheses.

An example of incommensurability in theories of code-switching may be found by considering the different views held by Poplack and Myers-Scotton regarding the proper scope of a theory of code-switching (see e.g. Poplack & Meechan, 1998, Myers-Scotton, 2002) vs borrowing. Here the problem of incommensurability arises because the notion of linguistic integration is key to the definition of borrowing for Poplack, while it is at best a hypothesis about borrowing for Myers-Scotton. We attempt a solution to this problem by critically examining the notion of linguistic integration in order to determine whether a clear line can be drawn between integrated and unintegrated donor-language items. The data we have used for this are English-origin verbs in data collected from Welsh-English bilingual speakers who speak mainly Welsh. We have subjected these to three tests of linguistic integration to see whether a clearcut distinction can be drawn between switches and borrowings. We show that the three tests have different results, and that the notion of a continuum between switches and borrowings is more defensible. Finally, we propose a new hypothesis to be examined in relation to the data, that the linguistic integration of donor-language items will be related to their frequency.

One of the interesting phenomena in language is that one word (or phrase) can introduce a syntactic commitment for the occurrence of a word or phrase with particular syntactic characteristics which can occur much later in the sentence. WH-phrases are one of the most studied of these dependencies. These are particularly interesting because the commitment is for a missing element (trace or gap). That is the WH-phrase Which boy in Which boy did John tell Susan that he went to the movies with ___ yesterday? has to be paired with an unfilled NP position like that following with; note that without the WH-phrase this sentence would be ungrammatical in most varieties of English. Research using ERPs has shown that WH-phrases introduce a memory load which is carried until the commitment is filled and that there are also effects at the point at which the gap is located which are modulated by the distance over which integration with the WH-phrase must extend. There are a number of interesting issues about the processing of long-distance dependencies. First, it has not been clear whether there are specific processing routines for WH-dependencies, or if similar effects can be found for other types of syntactic commitments. I will discuss an experiment that involves the processing of the particle zai in Chinese, which introduces a commitment for a locative postposition. Compared to sentences with no specific commitment (copula constructions), these sentences show a sustained negativity similar to that found for WH-sentences. There are also signs of costs of integration across distance which are similar to those found in WH-constructions. This suggests that these processes are not specific to gap location and filling, but reflect more general processes regarding maintaining and resolving commitments. A second issue has to do with the extent to which the processing effects described above should be considered to be those of syntactic commitment and resolution or of semantic commitment and integration. This can be addressed manipulating the degree of semantic commitment that is embodied in the word or phrase which introduces the long distance commitment. For example, Chinese classifiers are similar to grammatical gender systems in that they introduce a commitment for a particular type of head noun, but it appears to be much more semantic in nature than the syntactic commitment introduced by grammatical gender. Nevertheless distance to the point of integration induces a positivity which is similar to that found for the zai construction, in which the semantic constraint is considerably less detailed. The primary difference is that the effect is much larger for the classifier commitments. Likewise, manipulating the degree of semantic constraint of a WH phrase modulates the size of the maintenance effect over intervening material. These results suggest that the semantic aspect of the commitment may be as important as the syntactic aspects in the brain processes which are reflected in these two ERP effects.

Bilingual chlidren, as well as older adults, exhibit advantages over their monolingul peers in tasks that rely on executive control. However, until recently, studies comparing bilingual and monolingual college students found mixed results, and a less consistent bilingual advantage. Most studies examining this population have used tasks that rely on spatial visual attention, such as variations of the Simon task, the ANT task and the anti-saccade task. In this talk, I will describe a new study that compared the performance of monolingual and bilingual college students on three executive control tasks, and investigated possible bilingual advantages beyond the domain of spatial attention. Possible implications of the results for the locus of the bilingual executive advantage will be explored.

Words provide the central interface between form and meaning during language comprehension. Describing the nature of form-meaning interactions at the level of individual words is therefore one of the major goals of contemporary research on language comprehension. Part of that general endeavor involves describing exactly when semantic information becomes available during visual word recognition, and the nature of the form-level processing that is necessary for that to occur. I'd like to present some elements of response to these specific questions as well as address the question of the interrelation of the two languages of language learners at the word level. I will present a study using event-related potentials (ERPs) to examine the time-course of visual word recognition in second language learners using a masked repetition priming paradigm as well as other data from bilingual studies run in our lab that may shed light on these topics.

Children with specific language impairment (SLI) often show an uneven profile within the area of morphosyntax. For example, in English, the use of tense/agreement morphemes stands out as an area of special weakness. In Swedish, both word order and the use of tense can be problematic. These weaknesses are resolved only gradually. Thus far, the theoretical frameworks that might account for the findings constitute only partial solutions. Some provide a very insightful description of the difficulty but do not explain the systematic, incremental changes seen over time; others provide a plausible account of the gradual change but lack the precision necessary to explain the differences across languages. An alternative view that incorporates the empirically supported claims of the previous approaches will be offered. The alternative assumes that many of the characteristics of the SLI profile, including crosslinguistic differences in the profile, can be traced to details in the input, and that children's ability to interpret successively larger grammatical units in input sentences can lead to the gradual, incremental changes seen in the children's morphosyntactic use. Observations supporting these assumptions will be provided, and their theoretical as well as clinical implications will be discussed.

Speech communication involves a chain of events that ideally aligns mental representations in the talker with those in the listener. Links in the chain can be "broken" at many points, particularly in cases where the talker and listener approach each other with non-optimally aligned linguistic sound systems (e.g. when they do not come from the same native language background) or when the listener's access to the speech signal may be blocked by a hearing impairment or the presence of background noise. I will present a series of studies that aimed to understand how talkers and listeners repair these breakdowns in order to achieve talker-listener alignment. The first study examined talker adaptation to the listener. Specifically, we conducted a series of acoustic-phonetic comparisons of "clear speech" across languages with various phonological structures. A second study focused on the other side of the talker-listener channel by examining listener adaptation to the talker. In particular, we investigated listener adaptation to foreign-accented speech. Both of these studies examined talker-listener adaptation under laboratory conditions in which the talker and listener did not interact directly. A third study examined talker-listener interactions under more natural conditions of spontaneous, dialogue recordings. In this study we examined communicative efficiency and phonetic convergence in English conversations between pairs of native English talkers and in conversations between one native and one non-native talker of English. Together, these studies build a picture of speech communication as a bidirectional process of talker-listener alignment even in the case of communication between interlocutors who do not share a "mother tongue."

Dr. Trueswell will present a series of eye tracking experiments that explore how visual-spatial attention is allocated during the perception of simple and complex events. Eye movements were recorded during a variety ofdifferent tasks, including event description, passive viewing, and thecomprehension of novel verbs (e.g., "Oh look! Mooping!"). The results show that there is a tight temporal (and sometimes causal) relationship between the allocation of visual-spatial attention and the rapid linguistic choices speakers make when describing events (linguistic choices that include Subject/Object assignment and manner vs. path description). Dr. Trueswell also shows that children as young as three years of age are sensitive to the characteristics of these speakers’ eye gaze patterns, and use them in conjunction with linguistic evidence to infer verb meaning.