Colin Bannard: The place of multi word sequences in a productive language system
The last decade has seen a number of published findings that our competence (speed, fluency etc.) with word sequences reflects their frequency, supporting the suggestion that we have dedicated representations for sequences that we use or encounter a lot. This raises the challenge of explaining how such representations can fit into a productive system for word combination. In this talk I will present a series of studies that start to address this, looking at how frequency-sensitive processing of phrases interacts with other pressures on language use.
Denisa Bordag: Theories of second language acquisition: An overview
In the four sessions, selected theories and models of second language acquisitions will be reviewed chronologically, starting with the Contrastive Analysis Hypothesis of 1950’s up to e.g. Ullman’s Declarative/Procedural Model. Implications of the models for language teaching and learning will be discussed as well.
Harald Clahsen: Grammatical constraints in language learners’ spoken language comprehension
The human language processor is capable of rapidly integrating grammatical information with information from other sources during reading or listening. Yet, little is known of how (child and adult) language learners make use of grammatical cues along with other information sources during language comprehension. For the present talk, I will address this general question by investigating one specific phenomenon, the so-called plurals-in-compounds effect, the avoidance of plurals inside compounds in English and other languages (e.g. *rats eater vs. rat eater), which is subject to both structural and non-structural constraints. I will present new evidence from both compound-internal modifiers in English and German (e.g., rote Paprika(s)fresser ‘red pepper(s) eater’). Results will be reported from graded linguistic judgements and from visual world eye-movement experiments focusing on the role of grammatical constraints in both child and adult language learners’ spoken language comprehension.
Harald Clahsen: How the brain plans complex words: Silent production of inflected words during EEG
In this talk, I will examine brain potentials during the production of morphologically complex words. To study neural correlates of producing morphologically complex words, event-related brain potentials (EPRs) were recorded during silent production. I will discuss findings from a number of studies using this technique investigating different kinds of morphological phenomena in English and German across different populations, including native speakers (both adults and children) as well as and non-native speakers.
The results were found to be consistent across languages, phenomena, and populations. Rule-based forms elicited a characteristic pattern of brain potentials during (silent) production, which included an enhanced negativity relative to control conditions. I argue that this ERP effect reflects control processes involved in constraining rule-based computation during morphological encoding and discuss how this mechanism could be incorporated into models of language production. Finally, I will speculate that this particular ERP effect may belong to a family of negativities signalling cognitive control monitoring more generally.
Filip Děchtěrenko: Psychopy: How to build psycholinguistic experiments?
An open-source software for building experiments will be presented in the workshop. We will build a simple experiment, learn how to present texts or pictures and how to gather response information (using a keyboard, or a microphone). In the end, we will look into some experiments created using this software. It is not necessary to know how to do programming to take part in the workshop (but it could be an advantage for you). Please, bring your own laptop with an installed Psychopy version 1.84.2 (you may download it here: http://psychopy.org/installation.html).
Natalia Levshina: How to see the meaning: Exploration and visualization of semantics with R
This practically oriented course demonstrates how one can use multivariate statistical methods and graphical tools in order to visualize, explore and compare semantic structures, semantic relationships and semantic changes in one, two or several languages. We will discuss different types of semantic maps created with the help of Multidimensional Scaling, Multiple Correspondence Analysis and Graph-Theoretic algorithms. The case studies are based on typological data, parallel and comparable corpora. In addition, visualization and exploration methods for distributional semantics (Behavioural Profiles and Semantic Vector Space Models) will be presented, as well as ways of representing semantic change in time.
Magdalena Łuniewska: How to measure vocabulary size cross-linguistically?
As lexicon is the most language-specific domain of language, it is impossible to translate tests of vocabulary directly from one language to another. On the other hand, there is a need for comparable measures of lexical knowledge in both languages of a bilingual child. Typically, tools designed to identify SLI do not take into account if a child is bilingual and how this might affect raw test scores, often leading to misdiagnosis. Both vocabulary size and processing speed can be confounding variables when diagnosticians attempt to disentangle bilingualism from SLI at the lexical level. Lexical abilities can also be used as a baseline assessment of bilingual dominance/proficiency.
To address this need, we designed The Cross-linguistic Lexical Tasks (CLTs). CLTs provide a fully comparable assessment for vocabulary and lexical processing in 26 different languages. I will present the innovative method of the CLTs´ construction: a multilingual parallel task construction procedure which enables an objective test of vocabulary and processing skills in any pair of languages included in the process.
Magdalena Łuniewska: Mental lexicon of Polish SLI children
The issue of shortage of vocabulary among children with specific language impairment (SLI) is still disputed. Some researchers underline the grammar-specific nature of SLI, whereas others notice its signs also in mental lexicon. In the talk, I will present a study in which we examined whether Polish preschool children at risk of SLI present deficits in vocabulary size and in lexical processing. To answer this question, we applied three different measures of vocabulary: standardized test of picture-choice, picture naming and a scene-description task, as well as a measure of lexical processing based on reaction times in CLTs.
In the talk, I will contrast the outcomes of the three measures of vocabulary and lexical processing. In addition to the outcomes of the SLI group and a pair-matched control group, I will present the potential usability of the three tasks for SLI diagnosis in Polish.
Danielle Matthews: How children learn to communicate: infancy
In infancy children discover that they can communicate with others about the external world and that they can do so using conventional language (words). I will talk about naturalistic and experimental evidence of how children make these qualitative steps forward in communicative ability.
Danielle Matthews: How children learn to communicate: preschool I–II
During the preschool years children come to learn which words to use in which context. I will talk about a range of factors that constrain how they do so and discuss learning processes that lead to ever more effective communication. In part 2 I will consider why we see such large individual differences in children’s communicative effectiveness.
Petar Milin: Pimping-up linguistics: Approaching language with behavioural experimentation and computational modelling
Over the last two decades linguistics has been steadily moving away from questions of what might be possible (competence) towards studies of what is probable (performance); with this has come a change in focus, away from exceptional examples, small samples and anecdotal illustrations, towards large-scale quantitative studies. Three preconditions had to be met for this step-change to take place: (a) the rise of corpus linguistics with large and accessible digitalized corpora, (b) the return of quantification to language studies, in line with the core assumptions of usage-based approaches, and (c) an emerging interest in neuro-behaviourally motivated experimentation due to the field’s aspiration to provide descriptions of language that are in line with what is known about the brain and mind from other disciplines.
In my workshops, I will present a hybrid methodological framework that combines behavioural experimentation with machine-learning simulations, which generate experimentally falsifiable hypotheses and provide a justification for observed behaviour. For this I will rely on Learning Theory, with special attention to its relevance for language learning. I will show how this geminate approach, that strengthens traditional experimentation, has the potential to move linguistic theory forward.
Using data from my own research on Slavonic languages, I will introduce a number of experimental paradigms, such as lexical decision (including a discussion of priming techniques) and naming, self-paced reading and eye-tracking. I will explain when the selected experimental paradigms are (not) useful, what they (do not) show and how their respective output data can be explored and analysed. Next, I will move on to discussing machine learning models such as Naive Discrimination Learning (NDL; e.g., Rescorla-Wagner model) and Kalman Filtering. I will demonstrate how computational modelling strengthens an experimental approach, and how it naturally converges with a corpus-based approach. I will also compare learning-based models with statistical parallels such as regression, a technique that is well-known in linguistics. Here, too, I will focus on explaining when machine learning techniques are (not) useful, what they (do not) show and how the data has to be prepared and processed.
Linguistics and psychology have proven their readiness to embrace each other’s epistemologies and methodologies. In fact, both disciplines have reached the maturity to work towards true interdisciplinarity and fuel a change across the language sciences.
Andriy Myachykov: Interplay between knowledge representations
The question of the interplay between the domain-general and the domain-specific cognitive systems is at the heart of current psychological and neuroscientific research. This general question can be further subdivided into at least two more detailed questions. First, do mental representations that belong to different knowledge domains share properties due to a representational overlap between them? Second, how do different domain-specific processes interface with each other by means of tapping into the same components of the general cognitive architecture (e.g., visual attention)? In my presentation, I will report recent evidence from our research group confirming the existence of both types of representational interplay. First, I will present data from cross-domain priming studies showing that similar organizational principles may underlie representations from different abstract-knowledge systems (e.g. mental arithmetic, time, spatial and emotional semantics). Second, I will report data showing how visual attention may subserve activation of distinct knowledge representations by acting as an interface between them.
Andriy Myachykov: The role of attention in sentence production
The world that we perceive and describe changes constantly. If we believe our descriptions of the world to be accurate and consistent, we must assume that the content and the structure of our individual sentences accurately and consistently reflect the world’s constantly changing nature. If so, a comprehensive production system need to model the sentence generation process taking into account this basic assumption: Words, their linear arrangement, and the structures they are inserted in must somehow reflect the corresponding parameters of the observed and described event. This system must include representation of salience as its integral component. This interplay involves constant, regular, and automatic mappings between elements of a visual scene with their varying salience and structural arrangement of the sentence constituents as well as the grammatical relations between them. In this interplay, perceptual input contributes initially to this mapping process by providing information for further conceptual and linguistic encoding. Importantly, this information is not processed in an unconstrained fashion; instead, it is systematically filtered, selected, and relayed based on a regular interface between the aspects of attention and their corresponding counterparts in the conceptual and linguistic structures. Bottom-up and top-down features of the interface include noticeability, importance, and relevance. As a result, linguistic output reflects in a regular way event’s conceptual organization including the attentional state of the speaker. This mapping between attentional focus and structural choice is a part of a more complex mapping mechanism that I will refer to as Cognition-Language Interface. Specifically, my talk will consider theoretical and empirical bases of a complex interplay between the speaker’s attentional state and the linguistic choices they make during sentence production.
Michael Ramscar: The discriminative nature of human communication
Traditional studies of language assume an atomistic model in which linguistic signals comprise discrete, minimal form elements associated with discrete, minimal elements of meaning. Since linguistic production has been seen to involve the composition of messages from an inventory of form elements, and linguistic comprehension the subsequent decomposition of these messages, researchers have focused on attempting to identify and classify these elements, and the lossless processes of composition and decomposition they support, a program that has raised more questions than answers, especially when it comes to the nature of form-meaning associations.
By contrast, behavioral and neuroscience research based on human and animal models has revealed that “associative learning” is a lossy, discriminative process. Learners acquire predictive understandings of their environments through competitive mechanisms that tune systems of internal cue representations to eliminate or reduce any uncertainty they promote. Critically, models of this process better fit empirical data when these cue representations do not map discretely onto the aspects of the environment learners come to discriminate. The first two talks in this series describe the basic principles of learning, and the empirical basis for the belief that human communication is subject to the constraints these principles impose. They describe how, from this perspective, languages should be seen as probabilistic communication systems that exhibit continuous variation within a multidimensional space of form-meaning contrasts.
This systematic picture of communication indicates that discrete descriptions of languages at an individual (psychological) or community (linguistic) level must necessarily be idealizations. Idealizations inevitably lose information, and the third talk describes how the development of a discriminative, information theoretic approach to language leads in turn to the appreciation of the vast array of socially evolved structure that serves to underpin human communication, and explains why the overly abstract models of language of the 20th inevitably led to this structure being ignored.
Finally, since humans are linguistic animals, we might expect insights from a successful theory of human communication to extend beyond linguistics: The final talk describes how the application of discrimination learning and linguistics can shed new light on our understanding of lifespan cognitive development, revealing that the idea of ‘healthy cognitive decline’ is a myth reflecting science’s failings when it comes to naturalizing the minds it studies.
1. Discrimination learning, development and language
This talk lays out the computational account of the brain’s basic learning mechanisms and describes the results of a series of experiments examining the relationship between learning processes and language. It then describes how learning models offer answers to two fundamental questions in the study of language, explaining how patterns of brain development can account for the uniquely human capacity for complex linguistic communication, and how the discriminative nature of learning allows us to formalize the relationship between linguistic form and linguistic meaning.
2. Meaning, morphology and discriminative communication
This talk describes how basic learning mechanisms can be used to shed light on morphology and morphological development. In particular, it describes how linguistic claims about the absence of ‘negative evidence’ in language learning are undermined by actual learning models, and explains how these models account for patterns of children’s morphological overegularization as well as making surprising and successful predictions in this domain. It describes the view of morphology that emerges from discrimination learning – in which forms contribute not only to the reduction of semantic uncertainty in structured ways, but also to the reduction of uncertainty about upcoming forms themselves – and applies this approach to shed light on two aspects of language that have long puzzled linguists: noun class systems (aka grammatical gender) and the semantics of personal names.
3. Discriminative contrasts and the structure and distribution of meaning
Information theory has shown that exponential distributions are beneficial to the design of efficient communication systems, because they are both optimal for coding purposes and memoryless. It was recently shown that Sinosphere family names are exponentially distributed, and this talk reveals how, consistent with this, the empirical name distributions in English are also exponential, such that the distributional structure of names appears to be universal to the world’s major languages. The talk will then describe how the other empirical distributions of English are also exponential – indicating that the Zipfian distributions long thought to play a functional role in language are an artifact of the mixing of these empirical distributions – and goes on to describe how these socially evolved structures serve to facilitate the discriminative processes of human communication.
4. Discriminative communication and lifespan development
Performance on a range of psychometric tests declines as we age, which is taken to show that cognitive information-processing capacities also decline across the lifespan. The final talk presents a series of analyses and experiments that all point to a different conclusion: that the patterns of slowing / “forgetting” – and non-slowing / “non-forgetting” – seen in healthy adults simply reflect the consequences of continual learning from the statistical distributions that typify much of human experience. Using simulations using discriminative learning models and large linguistic data samples, it shows how the patterns of test performance usually labeled as “decline” emerge in standard learning models as the range of knowledge that they acquire increases. Finally, using the developmental models described in talks 1 and 2, it describes a global model of cognitive function in which many of the changes in grey and white matter that are currently misinterpreted as “neural atrophy” can be seen as adaptations that serve to regulate the behavioral and metabolic requirements of mature brains.
Rachel Smith: Speech rhythm I
Does speech have true rhythm, and if so what is its nature? In recent years, controversies around the concept of speech rhythm have mainly concerned which acoustic metrics offer valid bases to categorise languages into rhythm classes, and latterly whether the concept of distinct rhythm classes is well founded at all. Most of the arguments have been made on the basis of cross-linguistic comparisons. This presentation surveys the major debates in the field, and looks at what within-language comparisons—specifically, a series of production and perception studies comparing aspects of timing and rhythm across several dialects of British English—can bring to the debate.
Rachel Smith: Speech rhythm II
Rhythm can have powerful effects upon listeners, as anyone who has ever felt the urge to dance to music knows. How do these effects arise in our bodies and brains? Do listeners entrain to speech rhythm in similar ways as they do to music? This presentation briefly surveys relevant background in music psychology and neuroscience, and presents the results of several behavioural experiments and an ongoing electrophysiological study on how listeners respond to the coordination of rhythm across turn transitions in spontaneous dialogue.
Shravan Vasishth: Introduction to Sentence Comprehension I-III
The annotations together with references may be found here: http://www.ling.uni-potsdam.de/~vasishth/courses/SummerSchoolLitomysl2017.html
Jordan Zlatev: Motivations and conventions in actual and non-actual motions semantics
Lecture 1: Actual motion semantics: semantic categories, mappings, and types of motion situations
Motion (event) semantics has been a popular field of semantic typology, but it has suffered from conflicting definitions of key concepts such as path, manner and “satellite”, as well as the notion of motion itself. I start by presenting a taxonomy of motion situations inspired by a combination of Vendlerian situation types and phenomenological analysis (Zlatev et al 2010), and a theory of spatial semantics with clearly defined concepts (Zlatev 2007; Blomberg 2014). This semantic framework is then applied in the analysis of descriptions of motion situations in Swedish, French and Thai, based on video clips (Blomberg 2014). The results confirm expected typological differences between the three languages, but also reveal some more problematic findings, calling for a return to a classical topic: the relations between what is “coded” (i.e. part of the conventional meaning of linguistic expressions) and what can be “inferred” from the context and general experience.
Lecture 2: Non-actual motion semantics: phenomenological motivations and linguistic conventions
If actual motion concerns changes in a figure’s relative position in space (as proposed in the first lecture), non-actual motion (Blomberg and Zlatev 2014, Blomberg 2015) concerns dynamic experiential structures directed to situations that lack such change. Such experiences, of which at least three different kinds can be distinguished (affordance perception, scanning and imagination) can then be argued to motivate the existence of non-actual motion sentences such as “The road crawls through the desert”. First, I show how such a framework has advantages to alternative ones relying on notions such as “mental simulation” and “fictive motion”. Then I review existing and ongoing research on eliciting such sentences in the same three languages described in the first lecture: Swedish, French and Thai. Again, some features of the descriptions in each language can be traced to typological properties of the languages, reminding of the need to distinguish between potentially universal experiences and conventional semantics. At the same time, certain kinds of pictorial stimuli (displaying figures that allow human locomotion from a first-person perspective) appear to elicit the most NAM-sentences in all three languages, suggesting the potency of prelinguistic motivations.
Lecture 3: Motion-emotion metaphors: universal motivations, language-specific conventions, and contextual negotiations
There in ongoing debate between Lakovian “conceptual metaphor theory” placing the locus of metaphor in (universal) cognition, and more discourse-oriented approaches, arguing that creative language use (as well as that of other semiotic systems, such as gesture and pictorial representations) needs to be given priority. To help resolve this, I first present a conceptual framework, inspired by the integral semantics of Eugenio Coseriu, that operates with three distinct levels of meaning: a universal level of bodily experiences, a conventional level of language-specific norms, and a situated level of discourse and context-specific norms. Within this, I analyze linguistic expressions that concern emotions, but use motion verbs and related constructions, such as “She fell in state of deep depression”: motion-emotion metaphors (Zlatev et al 2010). Extending the data to eight languages (English, Swedish, Finnish, Estonian, Spanish, Bulgarian, Thai and Mandarin), and systematizing the methodology, I show how certain cross-linguistic patterns can be traced to the universal level of pre-linguistic motivations, while many differences have to with cultural and linguistic norms that show both areal and genealogical patterns. I also discuss the fuzzy boundary between such expressions and the non-actual motion sentences analyzed in the second lecture.
References:
Blomberg, J. & Zlatev, J. (2014). Actual and non-actual motion: Why experimental semantics needs phenomenology (and vice versa). Phenomenology and Cognitive Sciences 13(3): 395-418.
Blomberg, J. (2014). Motion in Language and Experience. Actual and Non-actual Motion in Swedish, French and Thai. PhD dissertation, Lund University.
Blomberg, J. (2015). The expression of non-actual motion in Swedish, French and Thai. Cognitive Linguistics, 26(4), 657-696.
Zlatev, J. (2007). Spatial semantics. In H. Cuyckens & D. Geeraerts (eds.) The Oxford Handbook of Cognitive Linguistics, 318-350. Oxford: Oxford University Press.
Zlatev, J, Blomberg, J. & David, C. (2010). Translocation, language and the categorization of experience. In V. Evans and P Chilton (eds) Language, Cognition and Space: The State of the Art and New Directions, 389-418. London: Equinox.
Zlatev, J, Blomberg, J. & Magnusson, U. (2012). Metaphors and subjective experience: motion-emotion metaphors in English, Swedish, Bulgarian and Thai. In Foolen, Luedke, Racine & Zlatev (eds.), 423-450. Amsterdam: Benjamins.