The Diversity in Language, Culture & Cognition Colloquium series is a joint initiative of the Centre for Language Studies and Max Planck Institute for Psycholinguistics. It brings together the Meaning, Culture and Cognition, Languages in Contact, and Multimodal Language and Cognition groups, as well as the Language and Cognition department, for cutting-edge talks on diversity. It is co-organized by Ezgi Mamus and Alex Carstensen. For further information, please contact Alex Carstensen. Meetings take place weekly on Thursdays at 3.45pm at the MPI.
22 MARCH 2018
Variation in Shawi (Kawapanan): A Flux Approach
Northwestern Amazonia is a crucible of linguistic diversity. Many of South American so-called “isolate” languages can be found in this region, Shawi and Shiwilu being two of those languages. They are the only remnant members of the Kawapanan language family (Valenzuela, 2011; Rojas-Berscia & Nikulin, 2016; Rojas-Berscia, 2017). In this talk, I will provide a socio-historical survey of Shawi, looking both at modern social dynamics in Shawi communities that could inform on the historical past of the language, and at internal lexical/morphosyntactic evidence that was neglected following the comparative method. Denying the relatedness of Kawapanan internal features with other groups would be unprofitable. Therefore, I sketched a model I dub the Flux Approach, inspired by Johannes Schmidt’s Wave Theory, which beyond mere historical sound change diffusion patterns, deals with sociohistorical signals as well as their modern repercussion in the shaping of ongoing language change. Ultimately, the goal of this model is to provide a richer picture of the internal dynamics of language diversification in South America, avoiding as much as possible reification-derived Tree-Model accounts of language change.
08 MARCH 2018
Code-blending in child and adult bimodal bilinguals
It is well-known that the two languages of bilinguals are ‘on’ and influence each other in many ways. Bimodal bilinguals, who use a sign language and a spoken language, are no different in this. However, bimodal bilinguals take advantage of an option that is not possible for unimodal bilinguals: production of both a spoken language and a sign language simultaneously, known as code-blending. Code-blending can be compared with code-switching in terms of both sociolinguistic functions and practice, and in linguistic-based generalizations and constraints. In this presentation I will discuss data from our studies of code-blending in both children and adults, seeking to move toward an understanding of the linguistic properties involved.
01 MARCH 2018
Cross-linguistic priming in bilingual adults and children: An experience-based perspective
It is evident from many studies that bilingual language production and comprehension is influenced by cross-language interactions, such as code-switching and syntactic transfer. These cross-language interactions can be caused by multiple forms of linguistic experience, not only ‘long-term’ experience like a bilingual’s experience with both languages in daily life, but also ‘short-term’ experience like interactive alignment in dialogue and structural priming across languages (i.e., the tendency to re-use recently processed words and syntactic structures across languages). I will present mostly experimental but also corpus-based research on how such different sources of linguistic experience influence bilingual language use. In doing so, I focus on both code-switched and non-code-switched language use, arguing that the same mechanisms apply in both modes of speech. Although most of the data I present is based on bilingual adults, I will also discuss recent research ideas on how to investigate these issues in bilingual children.
22 FEBRUARY 2018
Inferring language change from a typological gap: The case of Bantu object marking
As linguists with the aim to discover the properties of languages and our language capacity, it is fascinating to find a gap in the logical possibilities of linguistic behaviour – gaps suggest that there are clear boundaries to what is and isn’t possible in our human linguistic systems. Looking at object marking in more than 50 Bantu languages (spoken in sub-Saharan Africa), I found just such a gap, which suggests a relation between the salience of objects and their behaviour in ditransitive clauses. In this talk I illustrate the object marking data in Bantu, and explain 1) how we can understand the impossibility of one of the four possible parameter settings, and 2) how this gap determines and informs us about language change.
1 FEBRUARY 2018
Engineering the statistics of the environment: How people shape their experience to promote learning
From an early age, children guide their acquisition of new information through selective attention and active inquiry. Word learning exemplifies a key challenge in this process of self-directed learning: knowing when to consolidate the familiar and when to absorb the new. I account for noun learning as a process of balancing these competing goals, and model semantic bootstrapping of rare words based on knowledge of common ones. I then extend this framework to explain how children and adults actively manipulate their environment for faster learning of both words and concepts. I identify desirable difficulties in the active learning process that can be leveraged to better promote self-directed learning on tablets and teaching in science museums. Finally, I will demonstrate an implementation of these principles in a collection of adaptive tablet-based games aimed at cross-cultural, curiosity-driven education that I have built and tested in Brazil, Tanzania, and the US.
25 JANUARY 2018
The lexical semantics of Melanesia and beyond. Explorations in deep-time linguistic areas through the lexicon
The distribution of variation in the domains of phonology and morpho-syntax are regularly used to establish areas of linguistic convergence. By contrast, little is known about the ways in which the structure of lexicon can detect processes of language convergence. In this talk, I explore the ways in which lexico semantics and lexical constructions can be used to define Linguistic Melanesia and areas of even deeper time-depths.
18 JANUARY 2018
Systematic mappings between semantic categories of concepts and types of iconic representations in the manual modality
An unprecedented number of empirical studies have shown that iconic gestures, those that mimic the sensorimotor attributes of a referent, contribute significantly to language acquisition, perception, and processing. However, it has not yet been documented whether there are generalisable principles of the mappings of concepts with different types of iconic strategies (i.e., modes of representation; Müller, 2013). In Study 1 we elicited silent gestures to explore the implementation of the different types of iconic representation (i.e., acting, representing, drawing, and personification) to express concepts across five semantic domains. In Study 2, we investigated the degree of meaning transparency (i.e., iconicity ratings) of the gestures elicited in Study 1. We found systematicity in the gestural form of 109 concepts across all participants with different types of iconicity aligning with specific semantic domains: acting was favoured for actions and (manipulable) objects, drawing for non-manipulable objects, and personification for animate entities. We also found that specific couplings between mode of representation and semantic domain were regarded as more transparent than others: acting yielded higher ratings for actions, representing for object-related concepts, personification for animate entities, and drawing for non-manipulable entities. These data show that there are generalised biases to represent certain concepts with systematic gestural forms and that the interpretation of the meaning of gestures is modulated by the interaction between type of iconic depiction and semantic domain. We argue that these mapping principles may extend to all forms of manual communication (gesture and sign).
07 DECEMBER 2017
Look, the puppy’s barking! Perceptual language in interactions involving children
Similar perceptual experience can be encoded in diverse ways both across and within languages (e.g., Majid & Levinson 2011). For example, while an act of perception may be expressed with a generic verb like listen, onomatopoeic words offer quite specific renditions of auditory phenomena, or, in some languages, non-visual sensory experience can be marked through a bound evidential morpheme. What is the work of these different form classes in language use? I explore this question through qualitative examination of data from English and from two languages of Papua New Guinea (Duna and Kaluli), with a focus on interactions involving children (main corpora as described by Demuth et al. 2006; Schieffelin 1990). These data give rich indications of the role of perceptual language in establishing, confirming or sustaining joint attention episodes in early conversations, and, potentially, of its more covert relevance to the development of rhetorical skill. However, different language structures and cultural settings may offer different ways to bring this about.
Demuth, K., J. Culbertson & J. Alter. 2006. Word-minimality, epenthesis, and coda licensing in the acquisition of English. Language & Speech 49: 137–174.
Majid, A. & Levinson, S. C. 2011. The senses in language and culture. Senses and Society 6(1): 5-18.
Schieffelin, B. B. 1990. The give and take of everyday life: Language socialization of Kaluli children. Cambridge: Cambridge University Press.
29 NOVEMBER 2017
Towards a semantic typology of complex predicates: Auxiliary verb constructions vs. serial verb constructions vs. light verb constructions
The terms serial verb, auxiliary verb and light verb have each been used by various authors to cover virtually every complex predicate subtype and it is not clear what a given scholar intends when using them. To put some order into this chaos, I offer a semantic approach to distinguish them across all of the different language groups and traditions. First all such terms are meaningful only in a larger constructional context. Such formations minimally consist of two elements, one of which is purely lexical. The other element in such constructional frames can be designated the serial verb, auxiliary verb or the light verb. I propose that an element is an auxiliary verb if it contributes functional semantics to the overall constructional sense, a serial verb if it contributes lexical semantics, and a light verb if it contributes no other semantics except instantiating valence and/or serving as an inflectable verbal element in a complex predicate, the other component of which has predicative potential but cannot be inflected as a verbal element in the language for a variety of different reasons (it’s an ideophone, a borrowed stem, phonologically defective, etc.). Examples from many languages are used to demonstrate these claims.
23 NOVEMBER 2017
How learning to read changes mind and brain
Reading as a recent cultural invention has not been shaped by evolutionary processes and thus must make use of cognitive systems and brain networks which are either domain-general or have evolved for other purposes (cf. Dehaene & Cohen, 2007). Research on the effect of literacy thus is a powerful tool to investigate how cultural inventions impact on cognition and brain functioning. During my talk I will draw on evidence from both behavioural experiments and neurobiological studies.
I will present the results of a series of studies in which we found that illiterates, less proficient young readers, and people with dyslexia show similar language prediction deficits, with evidence consistent with the notion that this at least partly reflects their common reduced (or suboptimal) reading exposure rather than a causal impairment due to a reading disorder. I will also show that that (il)literacy has important consequences for the cognitive ability of selecting relevant information from a visual display of non-linguistic material. I will present two experiments which show that learning to read results in an extension of the functional visual field from the fovea to parafoveal areas, combined with some asymmetry in scan pattern influenced by the reading direction, both of which also influence other (e.g. non-linguistic) tasks such as visual search.
Finally, I will present the results of a longitudinal study with completely illiterate participants, in which we measured brain responses to speech, text, and other categories of visual stimuli with fMRI (as well as resting state activity and structural brain differences) before and after a group of illiterate participants in India completed a literacy training program in which they learned to read and write Devanagari script. A literate and an illiterate no-training control group were matched to the training group in terms of socioeconomic background and were recruited from the same societal community in two villages of a rural area near Lucknow, India. This design permitted investigating effects of literacy cross-sectionally across groups before training (N=86) as well as longitudinally (training group N=25). Our findings crucially complement current neurobiological concepts of normal and impaired literacy acquisition and highlight the need for the inclusion of diverse participant populations in psychological and neurobiological research.
16 NOVEMBER 2017
Early language experience in a Mayan village
Recent work on socioeconomically diverse Western samples has convincingly shown that children’s linguistic input affects their language learning. However, generally speaking, acquisition is quite robust. Children grow up to be competent language users under radically different conditions of caregiver-infant interaction across the world’s cultures. By studying development in non-Western cultures, we stand to better understand the scope of variation in early language experience—a crucial step toward finding out which aspects of language learning are sensitive to early experience and which are more robust. I will present initial findings from an ongoing project on linguistic and communicative development in children under age five from two non-Western cultures: one in which caregivers engage children in intensive, face-to- face verbal interaction from infancy (Papua New Guinean: Rossel Island) and one in which early caregivers instead aim to keep their infants calm and quiet (Tseltal Mayan: Tenejapa, Chiapas, Mexico). I will primarily focus on the Mayan community, in which caregiver-infant interaction is reported to most widely differ from Western middle-class norms.
09 NOVEMBER 2017
Explaining lexical diversity in dialect data. The influence of concept features
Why do we only have a few words for the concept ‘sober’ (e.g. sober, abstinent), while ‘being drunk’ can be described by a myriad of expressions (e.g. intoxicated, hammered and I’m not as think as you drunk I am)? Lexical variation of this type is often considered to be influenced by lectal features, like sociolinguistic or register-related stratification. For instance, in colloquial speech, a language user may refer to a drunk person as a someone who is hammered, while a doctor’s report is more likely to contain the term intoxicated. However, in this presentation, I will show that, in Dutch dialect data, variation in the amount of lexical diversity, i.e. the number of alternative expressions available to refer to a particular concept, is also influenced by the meaning of the concepts to be expressed. For tabooed concepts, like ‘being drunk’, a large degree of lexical diversity is found because names for these concepts quickly lose their euphemistic reading (Allan & Burridge 2006). Additionally, I show that features related to a maximalist, Cognitive Linguistic view on meaning interact with diversity as well. In the dialects of Dutch, concepts that are psychologically more entrenched, for instance, occur with significantly less lexical variation. Furthermore, I present evidence that differences in the amount of lexical diversity between groups of dialect speakers can occur as well: lexical diversity is also influenced by socio-cultural factors, even in closely-related dialects of a single variety.
02 NOVEMBER 2017
Participant assignment to thematic roles in Tzeltal: Eye tracking evidence from sentence comprehension in a verb-initial language (Garrido)
Studies using eye-movements to investigate predictive processing (i.e., anticipation of upcoming
words) during sentence comprehension have mostly focused on subject-initial languages and have
found that thematic role knowledge is used to restrict the set of possible interpretations quickly. In
subject-initial languages, participant assignment to thematic roles might be difficult to assess
because one argument is always available from the outset (often the Agent) but usually needs the
verb to be interpreted semantically. Therefore, it is not clear what type of information
comprehenders actually use to anticipate upcoming event participants in these languages. Verb-
initial languages provide the opportunity to investigate the early interpretation of events and the
type of information that drives anticipatory processing. I will present a visual world eye tracking
study on Tzeltal, a Mayan language spoken in Chiapas (Mexico), where I investigated how
information provided by verbs is used to assign participants to thematic roles. Tzeltal allows us to
test whether anticipatory eye movements to agents and patients are driven by 1) verbal semantics,
2) voice marking and word order (active: VPA or passive: VAP), or 3) if listeners follow a
(potentially universal) Agent preference. I will show that 1) verbal semantics alone is not sufficient
to guide visual attention towards referents in Tzeltal, 2) voice marking drives anticipatory
processing towards agent and patient referents (in passive sentences), and 3) there is no evidence
for a preference towards early agent fixations, challenging the universal salience of agents in
The role of intonation contours in turn-taking behaviour (Valtersson)
Turn transitions in everyday conversation are very rapid, with an average duration of 200
milliseconds across a wide variety of languages (Stivers et al., 2009; Heldner & Edlund, 2010). It
has been claimed that interlocutors are able to coordinate turn transitions with such precise
timing since addressees can plan their upcoming turn while the current one is unfolding
(Levinson & Torreira, 2015). For such purposes, it has been suggested that addressees can
identify the speech act type (e.g., a question) of the current turn early and prepare their upcoming
response accordingly (e.g., an answer). But such information only helps addressees to prepare
what to say. In this presentation, I present a study that addresses how addressees know when to
start speaking. Prosodic turn-final cues have previously been categorised as a ‘go-signal’ for
addressees to launch their pre-planned upcoming turn (Levinson & Torreira, 2015). I examine
how different types of intonation contours (i.e., the phrase-final melodic movements) by the
current speaker relate to the subsequent turn-taking behaviour in a corpus of spontaneous French
conversations. The results suggest that some intonation contours are used to hold the turn,
whereas others do not appear to influence the subsequent turn-taking behaviour. Since corpus
research provides correlational results, I also examined the causal effect of the type of intonation
contour in a judgment task with naïve participants. I phonetically manipulated utterances to
create two versions of an utterance with the same lexical information but different phrase-final
intonation contour. The task revealed that the type of intonation contour alone affects naïve
participants judgments regarding who will speak next.
26 OCTOBER 2017
Event construal under a cross-linguistic and acquisitional perspective (Von Stutterheim)
Cross-linguistic studies of the verbalisation of motion events which include Semitic (Algerian Arabic, Modern Standard Arabic), Germanic (English, German, Dutch, Norwegian), and Romance languages (French, Italian, Spanish) reveal language-specific effects in the way events are construed, showing that the underlying principles are both perspective driven and linked to patterns of grammaticisation.
The presentation will focus on French, Tunisian, German and two different L2s (Modern Standard Arabic and German). Speakers of these languages were shown the same set of short dynamic scenes as stimulus material. Data come from language production tasks and eye tracking. The linguistic data are analysed with respect to
• cross-linguistic differences in the selection of components representing events
• cross-linguistic differences in syntactic packaging
• the interaction between temporal and spatial categories
Results will be discussed in the context of spatial typology and in relation to theories of conceptual reorganisation in second language acquisition.
Predicting objects states – The case of the Mandarin Chinese bǎ-construction (Gerwien)
The ‘skeleton’ of an event representation can be thought of as an array of (multiple) time
spans, each of which binds one entity and one quality (Klein 2000). This view implies a
differentiation between a structural (the array of time spans) and a content level (the qualities
linked to entities at the time spans) of event representation. The Mandarin Chinese bǎ-
construction presents itself as an interesting case to evaluate this theoretical concept:
Besides other functions, the marker bǎ changes the canonical word order S-V- O to S-bǎ-O-
V; it marks the noun which it precedes as the direct object (cf. Yang & van Bergen 2007),
and it signals that the corresponding referent must be interpreted as having changed from
one state to another (cf. Li & Thompson 1981). The qualities of the referent’s initial and
resultant state, however, are specified by the sentence-final verb. Because in online
comprehension people use the current linguistic input incrementally to predict upcoming
discourse (cf. Altmann & Mirković 2009), it can be hypothesized that the function word bǎ
triggers predictions about the referent following it: bǎ activates an abstract, that is, a
qualitatively unspecified representation for an affected object in the comprehender’s situation
model and this representation interacts with the visual input, which leads to predictions about
upcoming linguistic input. I report results from two visual world studies which confirm this
19 OCTOBER 2017
Symbolic play provides a rich socio-cognitive context for language acquisition
There has been a long-attested empirical relationship between symbolic (or pretend) play and language acquisition. Early explanations of this relationship appealed to the inherently symbolic nature of the two systems. For instance, object substitutions such as pretending a block is a car mimics the mostly arbitrary symbolic relationship between words and referents. Following Piaget (1962), the play-language relationship was therefore attributed to the development of a general capacity for symbolic representation during later stages of sensorimotor development. In contrast, socio-cultural approaches to development (e.g., Rakoczy, 2008; Tomasello, 1999; Vygotsky, 1962) assume that the context of symbolic play provides a unique social context for development by virtue of both its symbolic nature (X represents Y in context C) and the fact successful play interactions require the establishment of collective intentionality. In this talk I describe a research project that re-visits the play-language relationship in light of suggestions from socio-cultural theory that social processes underlie the relationship.
I first present a meta-analysis that establishes the strength of the play-language relationship across development. I will then present the results from a longitudinal study that investigated socio-communicative behaviours in infant-caregiver dyads across two play contexts (symbolic versus functional play). We found that, compared to functional play, infant-caregiver dyads engaged in greater amounts of joint attention during symbolic play, and produced more representational gestures. Linguistic interaction also differed, with caregivers notably using more interrogatives and less imperatives in symbolic than in functional play. Finally, there were significantly more conversational turns during symbolic play, which longitudinally predicted children’s language development. The data suggest that symbolic play provides a fertile context in which infants can refine communicative skills that are important for language development.
12 OCTOBER 2017
Grammatical innovation in emerging signed languages: distinguishing alternate event construals
When new languages are formed, are some types of concepts grammaticalized before others? Observing how deaf individuals create new signed languages offers insight into this question. In my talk, I compare how adult signers of Nicaraguan Sign Language (NSL), adult Nicaraguan homesigners and child Guatemalan homesigners describe simple change-of-state events. “Homesign” refers to the gestural systems created by deaf individuals who do not have access to a sign language community but are not able to acquire the spoken language around them. Almost all of the individuals I studied use morphosyntactic devices to distinguish events with a causative agent (e.g., a woman tipping over a book) from events with no agent (e.g., a book falling over). Nonetheless, only those individuals who received signing input from an older peer used morphosyntactic devices to encode alternate construals of an event, i.e. viewing a causative event from the perspective of the patient rather than the agent. I discuss multiple explanations for this asymmetry, including the relative functional need to express each concept, as well as the possibility that these concepts require different levels of socio-cognitive maturity.
05 OCTOBER 2017
Unsung Heroes, Last Words, Elephants, and Other Tropes
To manage a flood of usable information and strategize about potential directions, journalists rely on tropes or story patterns. Most popular writing about language topics uses a restricted set of tropes, but many others could shed light on language in society and the language sciences more effectively. In this colloquium — a departure from the usual — I will present some common tropes as well as ones I regularly use (and ones I aspire to use more). This will serve as further introduction to my work and an opportunity to work with one trope, which could be called “Sketching the Elephant.” Here “the elephant” (which is a large topic, ungraspable as a whole) will be “the evolution of individual variation in linguistic production” or “the evolution of style.” Do individuals possess unique ways of speaking and/or writing? If so, why does this occur? People with expertise or insights about this topic are encouraged to attend. I will also describe the article I’m working on in which this discussion would fit.
28 SEPTEMBER 2017
Causality across languages: State of the art
This presentation introduces the project Causality Across Languages (CAL; NSF Award #BCS-1535846). The principal objective of CAL is the first large-scale comparison of how speakers of different languages categorize causal chains for the purposes of describing them. The presentation summarizes three ongoing studies, drawing on data from Datooga (Nilotic, Tanzania), English, Japanese, Korean, Mandarin, Russian, Sidaama (Cushitic, Ethiopia), Spanish, Urdu, and Yucatec (Mayan, Mexico and Belize). The first study targeted narrative production data. Results indicate remarkably parallel patterns of what is expressed (specifically or generically) and what is left implicit. The second study implemented a multiphasic design protocol combining production and the collection of goodness-of-fit judgments between descriptions and scenes. Regression models indicate a strong, highly significant correlation between domain (physical vs. psychological vs. speech act causation) and most compact construction type acceptable. Surprisingly, however, the involvement of an intermediate human participant did not exert a significant effect. The third study investigates the role of intentionality in attributions of responsibility. Participants watched scenes featuring two actors involved in a causal chain initiated by one of them. They then divided tokens into piles indicating their assignment of responsibility for the resulting event. A linear mixed effects regression model indicated a significant interaction between intentionality and population. This supports previous findings suggesting that the role of internal dispositions in responsibility attribution is variable across cultures. A follow-up question we intend to take up is whether folk theories of agency influence grammars.