Sonja Eisenbeiss

you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at

Most studies on bilingual children’s metalinguistic awareness (MA) assess MA using monolingual tasks. This may not reflect how a bilingual’s languages dynamically interact with each other in the formation of metalinguistic representations. We tested 33 Greek-Italian bilingual children (8-11 years) in their MA using acceptability-rating tasks in which they had to judge and explain grammatical errors. The tasks were in a monolingual and bilingual mode, in order to show how far MA in Italian benefited from the activation of Greek. Participants exhibited better MA abilities in Italian in the bilingual acceptability-rating task, in which Greek was activated. The benefits of the bilingual mode were visible in the judgement and explanation of errors and were modulated by syntactic processing abilities in Italian, length of exposure to Italian, type of structure and age. The results show that MA can be shared across languages. The pedagogical implications of the study are discussed.

This chapter provides an overview of theoretical issues and core empirical findings in cross-linguistic research on the acquisition of syntax. Section 1 identifies key issues in syntax acquisition research: (i) the respective contribution of learners' input and innate predispositions for language acquisition;(ii) the time course of syntactic development; (iii) the role of learners’ age and potential implications for monolingual, bilingual and second language (L2) acquisition. Section 2 introduces methods for investigating syntactic development. Section 3 discusses the relative role of learners’ input and innate predispositions for syntax acquisition. This section presents (i) generative, Optimality Theory and usage-based approaches to syntactic development and (ii) the empirical findings on learners’ input that form the background for the debate between proponents of the different approaches. Section 4 focuses on the emergence of syntax. The following sections discuss the acquis...

Of the total number of utterances (1231), the relatively high number of errors in Instrumental and Genitive Case markers (Table 1) are attributed to the child using an unfamiliar verb with a complex predicate structure for the first time as introduced by the researcher, and not because the child does not know the case marker. In the verbal elements (verbs and auxiliaries) children were observed to agree in gender with the object instead of with the subject even when there is no blocking of such agreement by overt case marking on the subject. Almost 40% of the agreement errors were observed to be due to this phenomenon. In order to

This article is published under a Creative Commons License CC-BY-NC (Attribution-NonCommercial). The licence permits users to use, reproduce, disseminate or display the article provided that the author is attributed as the original creator and that the reuse is restricted to non-commercial purposes i.e. research or educational use. See http://creativecommons.org/licenses/by-nc/4.0/ ______________________________________________________

Is motion cognition influenced by the large-scale typological patterns proposed in Talmy’s (2000) two-way distinction between verb-framed (V) and satellite-framed (S) languages? Previous studies investigating this question have been limited to comparing two or three languages at a time and have come to conflicting results. We present the largest cross-linguistic study on this question to date, drawing on data from nineteen genealogically diverse languages, all investigated in the same behavioral paradigm and using the same stimuli. After controlling for the different dependencies in the data by means of multilevel regression models, we find no evidence that S- vs. V-framing affects nonverbal categorization of motion events. At the same time, statistical simulations suggest that our study and previous work within the same behavioral paradigm suffer from insufficient statistical power. We discuss these findings in the light of the great variability between participants, which suggests...

The majority of experimental studies in linguistics, psychology, and the social sciences involve participants who are undergraduate students in research-active universities or children of educated families in societies that are WEIRD (Western, Educated, Industrialised, Rich, Democratic; see Henrichs et al. 2010). This leads to claims about universals of human language and behaviour that are not based on an appropriate empirical basis. It also violates the Principle of Justice as many populations are excluded from such studies and their benefits, for instance the development of appropriate materials for teaching or speech and language therapy. Hence, more and more experimental linguists have started to study previously under-researched languages and populations; and we are providing resources and information to support such projects (http://experimentalfieldlinguistics.wordpress.com/). These projects pose a broad range of ethical challenges. Some of them are challenges that any “traditional” linguistic fieldworker has to face, for instance avoiding coercion and guaranteeing informed consent when dealing with communities that are poor and characterised by low levels of education. However, the introduction of experimental methods into fieldwork contexts also gives rise to new ethical problems. In particular, we will discuss ethical issues that arise when standardized tests of performance (e.g. IQ-tests or tests of working memory) are carried out in small, close-knit communities where native speakers from the same community may become involved in the analysis of such data. We will also discuss (i) problems caused by data sharing in cross-linguistic and cross-cultural studies and (ii) conflicts that can occur when inter-disciplinary studies require ethical approval from boards with members from different disciplines (e.g. medical sciences, linguistics, psychology, and anthropology). Keywords: linguistics, justice, psycholinguistics, experiments, ethics

Im Bereich der Psycholinguistik zeichnet sich in den letzten Jahren eine zunehmende Kooperation zwischen theoretischer Linguistik und empirischer Spracherwerbsforschung ab. Aus psycholinguistischer Sicht hat sich die generative Grammatiktheorie nicht nur als Beschreibungsrahmen fur Spracherwerbsdaten, sondern auch als Heuristik zur Entdeckung von Entwicklungszusammenhangen bewahrt. Aus grammatiktheoretischer Perspektive erhofft man sich aus der Einbeziehung von Erwerbsdaten eine Erweiterung der empirischen Basis fur die linguistische Theoriebildung. Sollte sich die in den letzten Jahren zunehmend vertretene Auffassung bestatigen, das samtliche Ubergangsgrammatiken des Kindes in den Geltungsbereich der UG fallen, konnen nicht nur Daten aus der Erwachsenensprache, sondern auch Daten aus dem Erstspracherwerb zur Entscheidung zwischen konkurrierenden syntaktischen Analysen herangezogen werden (cf. Rizzi 1993).

This paper investigates the paradigmatic relations between inflected word forms (or their affixes) and the feature specifications of these elements. In two sentence-matching experiments German speakers had to decide whether sentence pairs involving inflected adjectives or determiners were identical or not. In both experiments, there was a delay when an inflected form contained positive feature specifications for grammatical features that did not match the feature specifications of the grammatical context in which it appeared. No delay, however, occurred when an incorrectly inflected form had mismatching negative specifications, whereas its positively specified features matched the respective positive features of the context. This result provides evidence for a different status of positively and negatively specified morphosyntactic features. It supports the idea of radical underspecification according to which only positive feature specifications are part of the representations of morphologically complex forms or affixes, whereas negative feature specifications are assigned on the basis of paradigmatic contrasts.

Sonja Eisenbeiss University of Essex [email protected] ... Dept. of Language and Linguistics, University of Essex, Wivenhoe Park, Colchester, Essex, UK, CO4 3SQ ... Essex Research Reports in Linguistics present ongoing research activi-ties of the members of the ...

Sonja Eisenbeiss University of Essex [email protected] ... Dept. of Language and Linguistics, University of Essex, Wivenhoe Park, Colchester, Essex, UK, CO4 3SQ ... Essex Research Reports in Linguistics present ongoing research activi-ties of the members of the ...

Collecting and analyzing samples of spoken speech from learners has played a central role in acquisition research since its beginnings. Initially, such speech samples were mostly collected in naturalistic settings, where researchers recorded learners in spontaneous interactions ...

The concept of 'event' has been posited as an ontological primitive in natural language semantics, yet relatively little research has explored patterns of event encoding. The study of Bhuvana Narasimhan and her team at the MPI for Psycholinguistic explored how adults and children describe placement events (e.g., putting a book on a table) in a range of different languages (Finnish, English, German, Russian, Hindi, Tzeltal Maya, Spanish, and Turkish). Results show that the eight languages grammatically encode placement events in two main ways, but further investigation reveals fine-grained crosslinguistic variation within each of the two groups. Children are sensitive to these finer-grained characteristics of the input language at an early age, but only when such features are perceptually salient. Our study demonstrates that a unitary notion of 'event' does not suffice to characterize complex but systematic patterns of event encoding crosslinguistically, and that chil...

This paper describes a coding scheme and a set of semi-automatic procedures for the annotation of complex noun phrases and their morpho-syntactic properties in child language data. These tools are based on the CHAT conventions of the Child Language Data Exchange System (MacWhinney 2000; CHILDES: http://childes.psy.cmu.edu/; CHAT: http://childes.psy.cmu.edu/manuals/chat.pdf). The coding scheme presented here focuses on the order and grammatical category of the individual elements in the noun phrase and their gender, number and case marking. It also provides information about the category and lexical identity of the element that assigns case to the respective noun phrase (e.g. the dative preposition mit ‘with’). The coding scheme was developed for German child language, but it can be adapted to other languages and populations.

"This paper discusses a series of so-called “elicitation” games that encourage children to talk in a situation that is as natural and relaxed as possible. Such games have played a central role in language teaching and speech therapy, where they have been employed to provide language training or to assess children’s linguistic development without putting them under stress. Recently, such games have become more widely used in language acquisition research. Here they are employed to obtain rich sets of language production data from children who are too young to take part in controlled experiments on language production. Moreover, they can be used in longitudinal studies where children are recorded over longer periods of time and might develop strategies in experiments. Most of these elicitation games target a specific construction or domain of grammar and so language teachers, speech therapists and researchers spend a lot of their time developing new games for each individual construction they would like to elicit from children. As this can be very time-consuming, there is a demand for games that can be adapted to a broad range of phenomena and situations (for instance, situations with one or several players). In this paper, I will present three such games, which might be useful for acquisition researchers, language teachers and speech therapists: the Bag Task, the Picture-Pairing Task and the Puzzle Task. In addition, I will discuss the advantages and disadvantages of using such games in a research context."

The main purpose of these reports is to provide a quick publication out- let. They have 'pre-publication status', and most will subsequently appear in revised form as research articles in professional journals or in edited books. ... Copyright remains with the author(s) of the reports. ...

Essex Research Reports in Linguistics present ongoing research activities of the members of the Department of Language and Linguistics. The main purpose of the reports is to provide a quick publication outlet. The reports have pre-publication status. The contents, form and distribution of the reports lie in the hands of the authors, and copyright remains with the author(s) of the

This paper investigates how morphological relationships between inflected word forms are represented in the mental lexicon focusing on paradigmatic relations between regularly inflected word forms and relationships between different stem forms of the same lexeme. We present results from a series of psycholinguistic experiments investigating German adjectives (which are inflected for case, number, and gender) and the so-called strong verbs of German which have different stem forms when inflected for person, number, tense or mood.

An Error Occurred Setting Your User Cookie. This site uses cookies to improve performance. If your browser does not accept cookies, you cannot view this site. Setting Your Browser to Accept Cookies. There are many reasons why a cookie could not be set correctly. ...

This chapter provides an overview of theoretical issues and core empirical findings in cross-linguistic research on the acquisition of syntax. Section 1 identifies key issues in syntax acquisition research: (i) the respective contribution of learners' input and innate predispositions for language acquisition;(ii) the time course of syntactic development; (iii) the role of learners’ age and potential implications for monolingual, bilingual and second language (L2) acquisition. Section 2 introduces methods for investigating syntactic development. Section 3 discusses the relative role of learners’ input and innate predispositions for syntax acquisition. This section presents (i) generative, Optimality Theory and usage-based approaches to syntactic development and (ii) the empirical findings on learners’ input that form the background for the debate between proponents of the different approaches. Section 4 focuses on the emergence of syntax. The following sections discuss the acquisition of core syntactic phenomena: questions and embedded clauses (section 5), passives (section 6), co-reference (section 7), and quantification (section 8). Each of these sections gives an overview of theoretical accounts and empirical findings; with a focus on monolingual first language (L1) acquisition. Age effects and differences between monolingual and bilingual acquisition are the focus of section 9. Section 10 discusses the empirical findings and their theoretical implications and highlights current trends.

Perceiving and talking about events taking place in the world around us is an essential part of our everyday life and crucial for social interaction with other human beings. Visual perception and language production are both involved in this complex cognitive behavior and have been investigated individually in
numerous empirical studies. Extensive models have been provided for both domains (see Hoffmann 2000; Levelt 1989, for overviews). But an integrative approach to the interface between vision and speaking, to “seeing for speaking”, is still lacking. Psycholinguists have only recently begun to experimentally investigate how visual encoding and linguistic encoding interact when we describe events1 and their protagonists or participants (see Henderson and Ferreira, 2004b). These studies have answered some, but raised many more general and specific questions:
 How does visual encoding of events evolve; how detailed are representations
of the visual world generated at various points during visual encoding?
 How is visual encoding linked to stages of linguistic encoding for speaking?
 Is the visual encoding of an event influenced by the linguistic task that subjects have to perform in experiments (e.g., describing scenes with full sentences vs. naming individual scene actors, and so on)?
 Is visual encoding influenced by the type of stimulus – in particular, are there
differences between line drawings and naturalistic stimuli?
 Does the encoding of (parts of) coherent scenes differ from the encoding of (parts of) scenes in which objects, animals or people do not interact in ways that could be straightforwardly interpreted as meaningful, coherent action?
In what follows, we present data from some of our own empirical studies that speak to such questions. To facilitate this, we first introduce the methods that we and others used to study visual and linguistic encoding during language production: eye-tracking experiments and brief presentation of visual displays. Next, we briefly present information on stages of visual and linguistic encoding and summarize data on interactions between visual and linguistic processing. All this serves as a background for a series of experiments in which we used different experimental tasks, stimuli and stimulus arrangements to investigate what kinds of visual representations speakers create, depending on how they describe (parts of) events.We used both line drawings and naturalistic pictures, presented either for ample time, or only very briefly. We asked our subjects either to fully describe the scenes, tomerely name the action or one protagonist, or even just to indicate the location of one of the protagonists by a button push. Scenes were coherent in some experiments and non-coherent in others: the coherent scenes involved objects, animals or people that interacted in ways that could be straightforwardly interpreted as meaningful events; in contrast,
the non-coherent scenes involved simple arrangements of unconnected entities (objects, animals, and people) that did not seem to interact with one another in any meaningful way. The data from these experiments, in combination with previous results, provide some preliminary answers to the questions above: brief presentations (<300 ms) of highly complex pictures (line drawings or photos of action scenes) allow speakers to activate speech production relevant information such as the role and identity of protagonists and whether or not protagonists engage in a meaningful interaction. The latter information can be recognized even if duration of presentations is as short as 100 ms, so extraction of event coherence seems to be quite an automatic process. Knowledge about the action depicted
in a scene poses a special case, as it might necessitate overt attention shifts towards regions that allow apprehending the relevant action. For instance, in order to distinguish whether a person is scratching or stroking an animal one has to look at the person’s hands. This is especially so if actions do not allow
a straightforward interpretation based on the distribution of protagonists in the spatial layout of the scene. In general, event complexity determined speech onset latencies, error rates and gaze patterns of upcoming utterances. Based on
such detailed visual representations of the apprehension phase, the very first eye movements in our eye-tracking experiments are then quite task-driven and allow a clear distinction between utterances in which the speaker is going to name, e.g., the agent or action of scenes. In some situations the very efficient
apprehension phase is accompanied by a preview phase, in which a part of the scene is fixated that is not named first in an utterance.We found repeatedlymore and longer gazes on action regions when speakers were going to produce full sentences in comparison to tasks where speakers had to name the protagonists in list format. We assume that the (overt or covert) encoding of action-relevant information is the crucial step in apprehending and describing action events, because it is the action that distinguishes an event from a static state.

In this document, we describe the transcription conventions for the Eisenbeiss German child language corpora.1 These conventions are based on the so-called CHAT transcription conventions of the Child Language Data Exchange System (MacWhinney 2000; CHILDES: http://childes.psy.cmu.edu/; CHAT: http://childes.psy.cmu.edu/manuals/chat.pdf). We have incorporated modifications and additions of CHAT for German that were suggested by Stephany and Bast (1999; http://childes.psy.cmu.edu/intro/stephany.pdf) and by Heike Behrens (2006; pc.; http://childes.psy.cmu.edu/manuals/07germanic.doc#_Ref131136188).

This paper presents a set of Case Elicitation Games and Stimuli (CEGS). The aim of this elicitation tool kit is to encourage speakers to produce a broad range of case-marked forms in a variety of different syntactic contexts, including subjects, direct and indirect objects, prepositional phrases and noun phrases that are not selected by a verb or preposition. The games involve two tasks - the Puzzle Task and the Picture-Pairing Task (Eisenbeiss 2009, 2010). Both tasks are semi-structured and involve flexible procedures and an informal interactional setting. The same target words are used in different games and syntactic contexts, which allows for cross-context and cross-method comparisons. CEGS was designed to provide rich semi- naturalistic speech samples of speakers from the age of two years. It can complement spontaneous speech sampling and controlled experiments on the use and comprehension of case marking; and the games are particularly effective for children that are too young to take part in controlled production experiments on case acquisition. The picture stimuli described in this paper were designed for studies involving German children, but we will discuss how tasks and stimuli can be adapted to other languages or adult participants, and to speech therapy or language documentation contexts.

This paper describes a coding scheme and a set of semi-automatic procedures for the annotation of complex noun phrases and their morpho-syntactic properties in child language data. These tools are based on the CHAT conventions of the Child Language Data Exchange System (MacWhinney 2000; CHILDES: http://childes.psy.cmu.edu/; CHAT: http://childes.psy.cmu.edu/manuals/chat.pdf). The coding scheme presented here focuses on the order and grammatical category of the individual elements in the noun phrase and their gender, number and case marking. It also provides information about the category and lexical identity of the element that assigns case to the respective noun phrase (e.g. the dative preposition mit ‘with’). The coding scheme was developed for German child language, but it can be adapted to other languages and populations.

All proponents of generative approaches to language learning argue that the syntactic knowledge which language learners acquire is underdetermined by the input. Therefore, they assume an innate language acquisition device which constrains the hypothesis space of children when they acquire their native language. However, it is still a matter of debate how general or domain-specific this acquisition mechanism is and whether it is fully available from the onset of language acquisition. This article provides an overview of the different answers that have been provided for these questions within generative linguistics. Moreover, it shows how the generative concept of “learning” has been applied to the acquisition of syntax, morphology, phonology and vocabulary, language processing, L2-acquisition, non-typical language development, creoles and language change. Finally, current developments, merits and problems of the generative approach to learning are discussed. The focus of this discussion will be on efforts to reduce assumptions about domain-specific innate predispositions for language learning.

Learning and talking about their own possessions and the possessions of their peers and caretakers plays a central role in children’s daily life. It is unsurprising then that relationships between possessors and their possessions are amongst the first relationships that children encode when they start to
string words together (see e.g. Brown 1973); and it is no wonder that many psycholinguists have made use of this rich data source to address questions about the mechanisms that drive children’s linguistic development.
However, most of the available studies of the acquisition of possessive constructions that we will discuss have investigated only one or two possession-encoding constructions in an individual language. Moreover, the focus
has typically not been on the encoding of the possessive relation itself, but on other aspects of the respective possessive construction. For instance, possessive -s markers in German and English (e.g. Susi-s Huhn ‘Sue’s chicken’) were analysed in studies that investigated whether the syntactic
categories of the target language were already present in early child grammars (e.g. Eisenbeiß 2000; Marinis 2002, 2003; Radford 1990). In these studies, possessive markers were simply treated as morpho-syntactic realisations
of syntactic categories; and semantic aspects were largely ignored. Similarly, possessive constructions with two-place verbs like "have" and "belong" were investigated in studies of the acquisition of syntax-semantic mappings, but these constructions were just treated as one type of twoargument
construction and not compared to other constructions encoding
possession (see e.g. Bowerman 1990, Pinker 1984). To our knowledge, no study has yet provided a comprehensive cross-linguistic overview that focuses on the different ways in which possessive relationships are encoded linguistically.
In order to fill this gap, we will provide a cross-linguistic overview of studies of children’s acquisition of the constructions that their target language employs to encode possession. In addition, we will present new data from German child language and child-directed speech, and discuss the implications for theoretical linguistics and language acquisition research. Our focus will be on three ways of encoding relationships between PRs and possessed entities (see Heine 1997; Baron, Herslund and Sørensen 2001 for
overviews):
– adnominal possession: Both Possessor (PR) and Possessum (PM) are
encoded within the same noun phrase (e.g. my/daddy’s chickens, the
chickens of our neighbours, …);
– predicative possession: The possessive relationship is encoded by a twoplace
predicate such as have, own or belong or by be (e.g. I have a dog.
The dog belongs to me. This dog is mine);
– “external possession”: the PR and the PM are realised as arguments of
a verb whose lexical meaning does not involve the notion of possession
(e.g. I tapped him-PR on the shoulder-PM).
We will first show how studies of children’s possession constructions can help us to evaluate models of children’s linguistic development. Against this background, we will present studies of the acquisition of adnominal, predicative and external possession constructions (EPCs). For each of these construction types, we will provide a brief overview of possession constructions in adult German and contrast them with possessive constructions in other languages for which acquisition studies are available. This will allow us to discuss empirical findings from earlier studies and our own analysis of German child data. Finally, we will compare the development of the three types of possession constructions and discuss the implications of our findings for theoretical linguistics and models of children’s linguistic development. In particular, we will show how the available empirical findings about the acquisition of possession constructions can be captured in
approaches that try to integrate core insights from current generative and usage-based approaches.

This paper argues for the integration of child language data into language documentations projects and shows the benefits that the documentation of child language can have for (i) acquisition researchers, (ii) descriptive, theoretical and historical linguists, and (iii) members of language maintenance or revitalisation projects. Moreover, it discusses which implications different user requirements have for data collection and provides pointers to resources, tools and stimuli.

Slobin, D., Bowerman, M., Brown, P., Eisenbeiss, S., Narasimhan, B. (2011) Putting Things in Places: Developmental Consequences of Linguistic Typology. In: J. Bohnemeyer & Eric Pederson (eds.), Event Representation in Language and Cognition. Cambridge: Cambridge University Press, 134-165.
"In this chapter, we explore how different languages describe events of putting things in places, and how children begin to talk about such events in their very early multi-word utterances. Our aim in focusing on the domain of “putting” events is to allow us to identify some important semantic and psycholinguistic factors that influence the course of acquisition. The overarching question is to determine the extent to which the development of linguistic event representations is influenced by the particular language the child is learning. Events of “putting” are frequently
discussed in interactions between caregivers and children, providing us with a rich crosslinguistic database in a high-frequency semantic domain. By examining language-specific characteristics of early event representations, we can make inferences about the cognitive resources and abilities that children bring to the task of learning how to talk about events in their native language.
A major motivation for working crosslinguistically is to investigate the role of language typology in children’s mapping of meanings onto forms—in this case, the expression of particular sorts of transitive motion events. In his well-known typology of how languages encode motion events, Talmy (1991, 2000) distinguishes between “satellite-framed” languages and “verbframed” languages on the basis of the element in the clause where information about path is
characteristically encoded. Our analyses show that this typological distinction does play an important role in the course of language acquisition, but other features that crosscut this typology play a role as well. These include properties of the target language’s inflectional morphology and
its semantic categories. We examine eight languages—four satellite-framed (English, German,Russian, Finnish) and four verb-framed (Spanish, Hindi, Turkish, Tzeltal)."

This paper discusses a series of so-called “elicitation” games that encourage children to
talk in a situation that is as natural and relaxed as possible. Such games have played a
central role in language teaching and speech therapy, where they have been employed to
provide language training or to assess children’s linguistic development without putting
them under stress. Recently, such games have become more widely used in language
acquisition research. Here they are employed to obtain rich sets of language production
data from children who are too young to take part in controlled experiments on language
production. Moreover, they can be used in longitudinal studies where children are
recorded over longer periods of time and might develop strategies in experiments. Most
of these elicitation games target a specific construction or domain of grammar and so
language teachers, speech therapists and researchers spend a lot of their time developing
new games for each individual construction they would like to elicit from children. As
this can be very time-consuming, there is a demand for games that can be adapted to a
broad range of phenomena and situations (for instance, situations with one or several
players). In this paper, I will present three such games, which might be useful for
acquisition researchers, language teachers and speech therapists: the Bag Task, the
Picture-Pairing Task and the Puzzle Task. In addition, I will discuss the advantages and
disadvantages of using such games in a research context.

This issue investigates the linguistic encoding of events with three or more participants from the perspectives of language typology and acquisition. Such “multiple-participant events” include (but are not limited to) any scenario involving at least three participants, typically encoded using transactional verbs like ‘give’ and ‘show’, placement verbs like ‘put’, and benefactive and applicative constructions like ‘do (something for someone)’, among others. There is considerable crosslinguistic and withinlanguage variation in how the participants (the Agent, Causer, Theme, Goal, Recipient, or Experiencer) and the subevents involved in multipleparticipant situations are encoded, both at the lexical and the constructional levels.

This paper investigates the paradigmatic relations between inflected word forms (or their affixes) and the feature specifications of these elements. In two sentence-matching experiments German speakers had to decide whether sentence pairs involving inflected adjectives or determiners were identical or not. In both experiments, there was a delay when an inflected form contained positive feature specifications for grammatical features that did not match the feature specifications of the grammatical context in which it
appeared. No delay, however, occurred when an incorrectly inflected form had mismatching negative specifications, whereas its positively specified features matched the respective positive features of the context. This result provides evidence for a different status of positively and negatively specified morphosyntactic features. It supports the idea of radical underspecification according to which only positive feature specifications are part of the representations of morphologically complex forms or affixes, whereas negative
feature specifications are assigned on the basis of paradigmatic contrasts.

In this study, we examined the system of case marking in two groups of German speaking children, 5 children with specific language impairment (SLI) and 5 typically developing (TD) children matched to the children with SLI on a general measure of language development. The data from both groups demonstrate high accuracy scores for structural case marking and over applications of structural cases to instances that require lexical case marking in the adult language. These results, we argue, provide evidence for the sensitivity of both TD children and children with SLI for abstract, structure-based regularities and are incompatible with accounts of SLI that posit broad syntactic deficits for these children.