SSoL 2016: lecture abstracts

Shanley Allen: The effect of discourse-pragmatics on referential choice in children I & II

One of the first major tasks children face in their linguistic development is learning to refer to objects. They must learn to choose appropriately between the various forms available such as lexical nouns, demonstratives, pronouns, and null forms. Many factors play a role in this choice including the pragmatics of both the discourse and the physical context. In these two lectures, we explore how children select referents in their speech consistent with pragmatic factors such as prior mention, perceptual availability, and joint attention. We investigate children’s sensitivity to these factors individually, incrementally, and in competition with each other. We then discuss reasons for this sensitivity, and their implications for theories of child language development. Data come mostly from English and Inuktitut, but cross-linguistic patterns are also discussed.

Shanley Allen: Cross-linguistic priming in adult bilinguals

Structural priming is the phenomenon by which a structure just heard or read influences you to produce the same structure in your subsequent utterance. If you have just heard the passive The apple was eaten by the boy (as opposed to the active The boy ate the apple), you are more likely to produce a passive in your next utterance. Much research with several structures and several languages shows that structural priming holds reliably within one language, and also often across languages. In this lecture, we first discuss the foundations and likely mechanisms of structural priming. We then explore one topic of current research in this area: what aspects of information does the formulator makes reference to when choosing between two alternative structures? Does it only attend to the elements which are part of the structure itself (local account), or does it attend to the properties of the entire sentence within which the structure is contained (global account)? Further, does this process work similarly within one language as it does across the two languages of a bilingual? We explore this on the basis of three studies of within- and across-language priming with German native speakers and German-English bilinguals. We conclude that formulation is not only affected by properties of the elements within the structure itself, by also by the global structural context, in both within- and across-language processing.

Shanley Allen: Developing tools for language assessment in Inuktitut (CDI, LARSP, calculating MLU)

Although there are many tools available for assessing language development in the world’s most-studied languages, such tools are still being designed and adapted for less-studied languages. In this lecture, we describe our experience so far in adapting tools for Inuktitut, the Inuit language spoken in arctic Canada. We specifically focus on ways that the morpho-syntactic structure of Inuktitut is different from that of English, and the implications this has for constructing linguistically appropriate tools. In particular, Inuktitut exhibits extensive agglutinative morphology, a rich nominal and verbal inflection system, frequent argument ellipsis, and polysynthesis. The tools include the Language Assessment Remediation and Screening Procedure (LARSP), the MacArthur-Bates Communicative Development Inventory (MB-CDI), and methods of calculating mean length of utterance (MLU).

Peter Bakker: Contact languages. An overview

All languages undergo influence from other languages, even supposedly isolated ones. Contact languages constitute extreme examples of linguistic influence. In normal transmission, languages are taken over wholesale with minor changes from previous generations of speakers. Contact languages on the other hand, have taken over so many elements from other languages, often within a short time period, that one can no longer claim that all the major parts have been inherited wholesale from one parent language. The most important of such languages are: (1) pidgins, severely reduced new versions of one other language, or a pair of languages, created in some situations in which no common language was available or attractive to the parties in contact. (2) creole languages, which inherited only part of the lexicon and grammar from the so-called lexifiers language (e.g. French, English), but their grammatical system is newly created and influenced by so-called substrate languages (e.g. West African languages). (3) mixed languages, which inherited systematic chunks from two different languages, e.g. lexicon versus grammar, noun phrases versus verb phrases, or overt morphemes versus abstract structure. 

In my presentation I will describe and exemplify these types of contact languages and link them with specific sociohistorical events. In my discussion, I will include the development of multi-ethnolects in current multilingual cities in Europe. Specific events are responsible for different types of contact effects.

Peter Bakker: Typology of creole languages

Recent research results have shown that creole languages are systematically different from non-creole languages in that they cover only a limited space of the typological possibilities of the languages of the world (Bakker et al 2011 in Journal of Pidgin and Creole Languages). This does not mean that creole languages are structurally homogenous: On the contrary, they are quite diverse structurally, especially if one takes also non-European creoles into consideration.

John McWhorter was one of the first to claim a creole profile of languages not found in non-creoles, with a set of three distinguishing properties, briefly: (virtually) no tone, no morphological tense and no non-transparent derivation, these being symptoms of young grammars. All three of these are based on absence of certain properties. 

More recently, mass comparisons using thousands of data points in computer programs have identified that the structural overlap between creoles and non-creoles is close to zero. Tests with large datasets always result in clusters of creoles, in which European languages sometimes participate. It appeared also possible to identify small sets of positive features (three, four or five) that set creoles apart from the non-creoles of the world. Thus, creoles have a shared typological profile, despite their diversity.

The most important results will be shown, not only from the Aarhus “Cognitive Creolistics” group, but also from other researchers. It will also be shown that less than a handful of structural features are sufficient to separate the world’s creoles from the world’s non-creoles, and that they can be used to locate formerly undetected creoles.

Peter Bakker: Mixed languages and other new languages

The status of mixed languages shows a turbulent history. Sometimes the existence of mixed languages was denied, sometimes all languages were considered mixed, albeit to a greater or lesser extent. 

One can define mixed languages as languages that have systematically inherited complete chunks from two identifiable languages. Under such a definition, the number of mixed languages is limited to a few dozen cases. Creole languages would not be mixed languages, as only the lexicon is clearly from one language whereas the sources of the grammatical structures are ambiguous.

Mixed languages can be divided into several types, based on what is taken over from the two source languages: (1) languages with a lexicon from one language and a grammatical system from another (G-L mixed languages, e.g. Media Lengua and a few dozen others), (2) languages with verbs from one language and nouns from another (V-N mixed languages, e.g. Michif, and a handful of others), (3) languages with verbs from one language and nouns from two languages (V-NN mixed languages, e.g. Gurindji Kriol), (4) metatypized languages (F-R mixed languages), in which all the roots for lexical and grammatical words are from one language, but the structures (frames) are copied from another language (no ideal exemplars, but Sri Lanka Malay, Takia and Kupwar languages come close). (4) languages with a lexicon from one language and the verbal inflection from another language (L-INFL mixed languages; Mednyj Aleut and one other known language). (5) mixed lexicon languages (L-L languages), all of which are pidgins or derived from pidgins (e.g. Berbice creole, Saramaccan creole). Few of these mixed languages, but far from all, have mixed phonological systems. 

I will contrast these mixed languages with other results of intensive contact, such as heavy lexical borrowing, codeswitching, secret and special languages, including languages spoken by settled nomads. 

Dan Dediu: Introduction to genetics for language scientists I–III

Studying language cannot be done in isolation, but must be well integrated in the wider landscape of modern science, of which genetics is an essential part. This series of three lectures aims to introduce some fundamental concepts, methods and results using a set of fascinating examples relevant to language and speech, aiming to provide a solid basis for understanding the primary literature and to be an active contributor to multi-disciplinary teams. Moreover, this overview should allow a more informed appreciation of the various proposals concerning the biological foundations of language, its origins and evolution.

Dan Dediu: Language does not evolve in void: anatomical and physiological biases in language change and evolution

It is becoming clear that language change and evolution does not happen in blissful isolation from the wider environment in which language is used, learned and processed, but that this environment shapes various properties of language and speech. In this lecture I will review the basic ideas and a few recent examples, focusing in particular on ongoing efforts to understand if — and how — inter-individual and cross-population variation in the anatomy of the vocal tract might influence phonetics and phonology, contributing to the emergence of linguistic diversity.

Tomáš Duběda: Intonation: from contours to tones    

Intonation is a universal property of speech that conveys both linguistic and paralinguistic information. Because of its great contextual, situational and individual variability, intonation has long resisted scientific formalisation. The first intonational models that emerged tried to capture sentence-level intonation holistically, while later proposals tried to structure the intonation phrase into units of different size. Tonal phonology, which has been in vogue since the 1980s, breaks down the melodic curve into a sequence of purely local events. Its recent application to Czech seems promising, or at least not disappointing: tonal stylisation offers a theoretically stimulating interpretation of the intonation system, and is back-compatible with the traditionally accepted contour analysis.

Josef Fulka: Deafness and Sign Language in Western Tradition

The aim of the lectures is to look at different ways in which sign language has been depicted and conceptualized in the history of Western thought. We will concentrate on two paradigms which appear to be present since antiquity up to the present time: deafness as deficiency (i.e. the absence of sensory perception and language) and deafness as difference (i.e. presence of a different mode of expression). We will show that the two different ways of looking at sign language clearly anticipate the conflicting views which are still present in contemporary debates on the issue in question.

Martin Haspelmath: Explaining universals of syntactic argument coding

In this presentation, I will discuss various ways in which the coding properties of verbal arguments have been explained by linguists, e.g. accusative and ergative marking, or presence or absence of person-indexing of objects. The focus will be on standard transitive and ditransitive constructions. I will distinguish between internal explanations, which are really the same as elegant descriptions, and functional-adaptive explanations, which make reference to language use and language change, and are necessarily explanations of language universals. I will evaluate a number of proposals that have been made recently and not so recently, putting emphasis on the form-frequency correspondence principle, which has wide-ranging explanatory power in diverse domains of grammar. I will argue that attempts to elevate internal explanations of particular phenomena to general principles of human language often fail because they make unwarranted (aprioristic) uniformity assumptions. Functional-adaptive explanations, by contrast, do not have to make such assumptions and are readily testable by cross-linguistic and usage data.

Ágnes Lukács: Natural Language Acquisition and Statistical Learning

The acquisition of complex motor, cognitive and social skills like language (or playing a musical instrument or mastering sports) is generally associated with implicit skill learning which often involves statistical learning (SL). In the two talks, we will focus on SL as a model of language acquisition in child language development, but we will also touch upon sequence learning and categorization. We will evaluate the role of SL in learning different aspects of language (phonology, lexicon and grammar), discuss the debate over language-specific versus domain general mechanism in language acquisition. Summarizing results both from previous studies and from our lab, we will look at how the development of SL mechanisms relates to the proposal of a critical period of language learning, and see how this mechanism operates in developmental and acquired disorders of language.

David Lukeš: The Python programming language: An introductory phrasebook for linguists

Why learn programming? – It’s useful for a linguist to know how text is represented inside a computer, how it can be collected in large amounts from the Internet and subsequently manipulated, what tools exist for its automated linguistic annotation. Natural Language Processing tasks like morphological tagging or syntactic parsing are fairly well understood today, and tapping into their power opens up research avenues you’d be unable to explore otherwise.

Beyond just text processing, any kind of empirical research requires data analysis, which in this day and age invariably implies using a computer. Spreadsheet software is great as long as your data and analysis fit on one screen, or a couple of screenfuls at the most. But even then it’s easy to get lost in the series of manual steps you’re performing, not to mention the impossibility of quickly re-applying them to new data.

By contrast, in a programming language, you write scripts that describe all the actions performed on the data, providing an easily reusable recipe for the analysis you’ve conducted. You’ll develop skills that will allow you to spot recurring patterns in your analyses, abstract them away, refine them over time and apply them effortlessly to any new data that might come by. Or use a library where someone has already defined these high-level techniques for you.

And if you favor introspection over empirical research, and formal grammars over corpora, why not try and test if your grammar ideas are formal enough for a mindlessly logical computer to use them correctly? (That Chomsky never did so himself should be considered a challenge, not an excuse.)

Most importantly though, programming enables you to transform tedious, repetitive tasks into fun, creative ones. Instead of doing something manually, find a clever way to automate it, re-use the procedure whenever needed, perhaps share it with other researchers…

Why Python? – It’s free, both as in “free beer” (no charge) and “free speech” (open source, free for anyone to modify, improve upon and extend). It features:

  • great functionality for text processing and data munging
  • a simple and easy-to-learn syntax that reads almost like English
  • straightforward semantics – understanding what any piece of code exactly does boils down to a few simple rules

We’ll be using:

  • the Anaconda Python 3 distribution – an easy way to get Python and some great third-party libraries and tools all at once
  • the Jupyter Notebook – a powerful data analysis environment for Python
  • the Natural Language Toolkit a.k.a. NLTK – a great Python library for linguists which comes with an even greater introductory textbook

Check these out in advance if you want to get a head start!

Asifa Majid: Semantic Typology: General introduction (lecture 1)

Provides the theoretical background to semantic typology from linguistics, psychology, and anthropology. Introduces the notion of a semantic domain. Introduces systematic stimulus-based elicitation studies in the field. Covers the methodological principles and challenges to semantic typology.

Asifa Majid: The language of perception (lecture 2)

Since Berlin & Kay (1969), researchers have been aware of both variation (different languages have different numbers of colour terms) and regularity in the semantic categories of colour (colour terms appear in languages in a specific order). Does this model hold up to modern-day scrutiny? And how well does it apply to other perceptual domains, like sound and smell?

Asifa Majid: Event semantics (lecture 3)

Events are considered ephemeral and therefore open to more relativity across languages than perception, for example. Are there, nevertheless, cognitive constraints at play in this domain? Results from human locomotion and separation events, amongst others, will be considered.

Barbora Skarabela: Is ‘choochoo’ better than ‘train’?: The role of baby-talk words in early language development 

The aim of this class is to examine the cross-linguistic characteristics of baby-talk words, their function and origin. In particular, we will focus on the hypothesis that the sound patterns of baby-talk words facilitate early lexical learning, and discuss this claim in light of the on-going debate on the role of early learning biases in adult language systems.

Virve-Anneli Vihman: Ways with words: Cross-linguistic acquisition of morphology

This course gives students an understanding of the theoretical issues at stake and methods used in research on the acquisition of morphology. Our focus will be on cross-linguistic variation and the different challenges linguistic systems pose for children. We will discuss approaches to investigating knowledge of morphology in early acquisition and the relevance of acquisition data to linguistic theory and developmental psychology.

Lukáš Zádrapa: Structural Pecularities of Old and Modern Chinese

In the course of the six hours allotted to the topic of linguistic description of Chinese, we will successively take up several issues that pose special difficulties in such undertaking. Because of the basic typological properties of the language, both ancient and modern, may serve as an acid test of general theories of language, that were often modelled on a Western language. Although in a broader comparative perspective, Chinese is not as exotic any more as many other tongues, there are many problems with application of even the most fundamental concepts. We will thus have to deal with such basic things as free/bound morpheme, affix, word – which is maybe the trickiest of all, the blurred boundary between morphology and syntax, but also with some typologically remarkable categories and constructions, such as measure words, disposal construction, serial verb constructions, or “modified verbs”. On top of that, elementary observations on the relationship between Chinese writing and Chinese language will have to be discussed.