University of Malta
 

Linguistics Circle 2018-2019
UOM Main Page
 
 
 
Apply - Admissions 2016
Newspoint
MSc in Language and Computation
BSc in Human Language Technology
Academic English Programme
Campus Map button

Linguistics Circle occasional seminar series

The Institute of Linguistics organises an occasional series of seminars, where local and international speakers are invited to present the results of recent research which is of interest to both professional (computational) linguists and members of the broader public.
The seminars take place in an informal setting and their primary aim is to encourage discussion. Students and members of the public are welcome to attend!

I am always looking for speakers in the Linguistics Circle Seminar. Please contact Lonneke van der Plas if you have something to present.

 

Current Seminars (2018-2019):

(Seminars from previous years:2016-20172015-20162014-20152013-20142012-20132011-20122010-2011)  

 ___________________________________________________________________________  

UPCOMING NOTICES:

 

Dr. Olga Bogolyubova, University of Malta

Title: Personality Traits and Subjective Wellbeing in the Language of Social Media

Date: Wednesday 26th June 12-14 

Venue: GW214

  

The notion that one can learn something about character and psychological functioning of an individual by studying how this person uses language has been around for some time and dates back to Freud’s interest in slips of the tongue. Today, when social media has become ubiquitous, psychological science can benefit from historically unprecedented access to rich natural language data. Existing studies demonstrate that language used online is associated with personality traits, emotional states, and even health conditions. This talk will present two studies based on the data from a sample of Russian adult Facebook users (n = 1,972). Data collection was conducted in 2015 via a purpose-built application, which served two purposes: it administered the survey and collected public wall posts from Facebook accounts of consenting study participants. The aim of Study 1 was to assess the connection between dark personality traits, harmful online behaviors, and language. The Short Dark Triad Scale was employed to explore a set of objectionable personality traits – Psychopathy, Machiavellianism, and Narcissism.  In Study 2, the focus was on the linguistic correlates of subjective wellbeing as measured by the WHO-5 instrument. In both studies participants’ wall posts were subjected to morphological, lexical, and semantic analyses. A number of significant linguistic correlates were identified for each of the psychological constructs under consideration. 

________________________________________________________________________________________________________________________________________________

   

PAST NOTICES:


 Prof. Jieun Kiaer, University of Oxford

Title: Argument and Particle Realisation: Syntactic or socio-pragmatic behaviours?

Date: Wednesday 17th October 13-14

Venue: OH122

 

In this talk, I argue that argument and particle realisation is pragmatically-driven rather than morpho-

syntactically driven. Often, in theoretical linguistics, it has been assumed that arguments and particles

in SOV, agglutinative languages can be “arbitrarily” omitted or dropped. The process behind choice

making – (i) whether to realise an argument or particle and if to choose to do so (ii) which one to

choose – has been largely overlooked in syntactic literature.


Wasow (2002), Philips (2003), Hawkins (1994, 2004, 2014) among many others aimed to show that

efficient structure building plays a crucial role in understanding human languages. Yet, often these

claims are only considered within processing literature rather than in mainstream syntactic literature.

This is because it is not easily tenable to incorporate resource-sensitive nature of syntactic behaviours

within a Chomskian linguistic framework which assumes native speakers’ access to “limitless”

linguistic creativity.


Based on processing-driven frameworks - Dynamic Syntax (DS, Cann et al 2005) and my recent

proposal on Pragmatic Syntax (Kiaer 2014), I aim to show how natural language structures are built to

maximize efficiency, expressivity and empathy (3Emodel). I argue that the choices surrounding

argument and particle realisations are driven by the above principles to make the utterance socio-

pragmatically adequate. I shall include data in this talk from SOV order, agglutinative languages such

as Korean, Japanese and Turkish.


Selected References

Cann R, R Kempson, L Marten (2005) The dynamics of language. Oxford: Elsevier.

Hawkins, J.A. (2004) Efficiency and Complexity in Grammars. Oxford University Press,

Kiaer, J (2014) Pragmatic Syntax, Bloomsbury, London, UK.

Wasow, T (2002). Postverbal Behavior. CSLI Publications.

Philips (2003) Linear order and constituency. Linguistic Inquiry, 34 (1), pp. 37–90. 

___________________________________________________________________________    

Mr. Andrei Zammit, University of Malta

Title: Maltese Dependency Parsing using Deep Learning Techniques

Date: Friday 2nd November 12.00 -14

Venue: OH105

Applications such as information retrieval and sentiment analysis depend on natural language processing tools. Dependency parsing is one of the tasks performed in NLP that analyses the grammatical structure of a sentence by determining the relationships between the words in a sentence. Whilst there are several parsers for many European languages, Maltese remains a low-resourced language and currently there is no parser for Maltese. This work investigates parsing of Maltese by using novel Deep Learning and bootstrapping techniques from multilingual sources, with the aim of contributing to the increase in computational resources for Maltese and also to dependency parsing. Results show an Unlabelled Attachment Score of 90% and Labelled Attachment Score of 86% when using a Quasi-Recurrent Neural Network (QRNN) with a bootstrapped data source of Maltese and other Romance languages. To our knowledge, this is the first time that a QRNN is applied to the task of dependency parsing. Thus, we report on the applicability of this technique for the task of dependency parsing in general.

___________________________________________________________________________    

Prof. Manfred Krug, University of Bamberg

Title: Language change in Maltese English and patterns of linguistic globalization

Date: Friday 9th November 12 -13

Venue: GW HD1

This paper draws primarily on data from Maltese English and, to a lesser extent, Puerto Rican English, British English and American mainland English. The data stem from questionnaire studies; the focus is on the lexicon.

In a first step, we will investigate linguistic choices made by different age groups in Malta, thereby identifying language change in progress. More specifically (though not unexpectedly), it will be seen that younger speakers in Malta generally portray a more globalized linguistic behaviour than older speakers. More surprisingly perhaps (given a very short diachronic time span), the data also suggest real-time changes, as globalization tendencies are more pronounced in recently elicited data from 2014 than in the first questionnaire studies from 2008 and 2009. Finally, the changes observed for Maltese English will be integrated into larger patterns of linguistic globalization. 

 ___________________________________________________________________________   

 Dr. Natalie Schembri, University of Malta

Title: Interview-based research in multilingual research situations: Recommendations for best
practice

Date: Friday 16th November 12.30 -14

Venue: OH105

Interview-based research in multilingual settings requires significantly increased research effort compared to monolingual situations. Researchers carrying out non-English interview-based research who publish in English need more time, research effort and financial means to translate as well as transcribe and code the data. With a view to understanding how this might affect the research process and in an attempt to establish a code of best practice optimizing the research effort in such situations, we investigate the views and experiences of four European researchers carrying out interview-based research in multilingual situations for a COST project. Data will be collected initially through a questionnaire followed by in-depth interviews. The study aims to raise an awareness of the issues involved in interview-based multilingual research and to contribute to establishing principles of best practices optimizing research time, effort and financial resources in multilingual research projects.

___________________________________________________________________________


 Dr. Teresa Lynn, University of Dublin

Title: Developing the Irish Dependency Treebank

Date: Friday 23rd November 12 -14

Venue: GW HD1

Syntactic parsing is concerned with the linguistic structural analysis of language in text. Statistical parsers are data-driven and rely on the availability of syntactically annotated corpora (known as treebanks) from which they learn patterns of syntax in a given language. Treebanks are costly in both terms of development time and skills required. For this reason, low-resourced languages often lack both treebanks and statistical parsers.

In this talk I will report on the development of the first Irish dependency treebank and syntactic parser. I will discuss the linguistic structures of the Irish language (a low-resourced language), and the motivation behind the design of the final dependency annotation scheme. I will also demonstrate how we examined methods such as Active Learning to semi-automate the treebank development. Through empirical methods, we will see the impact our treebank's size and content has on parsing accuracy for Irish.  I will also briefly discuss our work in cross-lingual studies through the use of a universal annotation scheme and our involvement in the Universal Dependencies Project.

 ___________________________________________________________________________

 

Prof. Holger Mitterer, University of Malta

Title: Glottal speech sounds in German:  phonological status and acquisition by late L2 learners

Date: Wednesday 5th December 12 -14

Venue: GW HC

 

The glottal stop (coded as ‘q’ in Maltese) is a speech sound that is a phoneme in Maltese but assumed to only be a boundary marker in German. In contrast with this point of view, results from a word recognition task show that the two are similar. Additionally, German /h/ also behaves like the glottal stop. This indicates that German has two glottal consonants: /h/ and glottal stop. This has implications for teaching German to late learners (L2 learners) with native languages that do not contain glottal consonants. Following the orthodox view, /h/ should be easy to acquire as it is perceptually distinctive from all other sounds of German and the native language of the L2 learners. With the glottal stop added, the /h/ and glottal stop represent a sound pair that should be difficult to distinguish. We therefore tested the acquisition of these consonants by Italian L2 learners of German in production and perception. In production, L2 learners performed moderately well with about 70% native-like productions. However, omissions (dropping /h/ or glottal stop) and exchanges (using /h/ were the glottal stop would have been appropriate) were observed for both target sounds with similar frequencies. In perception, listeners were able to hear a difference if asked explicitly but ignored the difference for word recognition. Again, glottal stop and /h/ behaved similarly. This is surprising given the head start that /h/ has over glottal stop: Listeners are constantly reminded of /h/ through orthographic coding and only /h/ is discussed in high-school education. These results suggest two conclusions: First, perception and production are skills that are acquired relatively separate in L2 acquisition. Secondly, explicit instructions on how to use a new sound in an L2 have little effect on actual attainment, which seems to be governed by implicit acquisition. This, in turn, means that natural phonological acquisition is still possible for late learners, even though they operate less efficiently after specialization for the first language(s).  


 ___________________________________________________________________________

  Dr. Pawel Rutkowski, University of Warsaw

Title: What corpus data can tell us about sign language grammar: The case of Polish Sign Language

Date: Wednesday 13th March 12-14

Venue: GW154

 

Dr  Rutkowski will give a presentation about the Polish Sign Language corpus project - the largest annotated sign language corpus in the world.

He will illustrate the linguistic uses of the corpus and give some examples of how the grammar of Polish Sign Language is analyzed on the basis of the PJM Corpus data.


 ___________________________________________________________________________

   Dr. Adrian Muscat, University of Malta

Title: "Can you please pick up the bag next to the armchair"

Date: Friday 10th May 2019  12-14 pm

Venue: GW 114 

The automatic detection of spatial relations between objects in an image or the machine interpretation of spatial relations in grounded language is useful in areas such as robotics, multi-modal image search, image description generation and image question answering. In this talk we review machine learning models that predict spatial relations in images. We cite work from cognitive and psycho-linguistics on how language structures space to motivate the challenges in automating the task of spatial relation detection. We review the architectures of the state-of-the-art machine learning models and discuss their limitations. We conclude with a look into future work.

 

 _____________________________________________________________________

 

 Prof. Adam Ussishkin, University of Arizona

Title: Roots, or consonants? On the early role of morphology in lexical access

Date: Friday 24th May 2019  12-14 pm

Venue: GW 114 


Words consist of a phoneme or letter sequence that maps onto meaning. Most prominent theories of both auditory and visual word recognition portray the recognition process as a connection between these units and a semantic level. However, there is a growing body of evidence in the priming literature suggesting that there is an additional, morphological level that mediates the recognition process. In morphologically linear languages like English, however, morphemes and letter or sound sequences are co-extensive, so the source of priming effects between related words could be due to simple phonological overlap as opposed to morphological overlap. In Semitic languages, though, the morphological structure of words reduces this confound, since morphemes are interdigitated in a non-linear fashion. Semitic words are typically composed of a discontiguous root (made up of three consonants) embedded in a word pattern specifying the vowels and the ordering between consonants and vowels. Active-passive pairs in Maltese illustrate this relationship (the root is underlined); e.g., fetah ‘open’-miftuh ‘opened’.

In this talk, I report on a series of experiments on Maltese investigating the extent to which root morphemes facilitate visual and auditory word recognition, and to what extent potential priming effects are independent of the phonological overlap typically inherent in morphological relationships. These experiments make use of the visual masked (Forster and Davis, 1984) and auditory masked (Kouider and Dupoux, 2005) priming techniques. The results of the experiments show that not only do roots facilitate visual and auditory word recognition in Maltese, but that these morphological effects are independent of phonological overlap effects.

 

_______________________________________________________________________________

 

 Prof. Andy Wedel, University of Arizona

Title: Words and phonological rules evolve to support efficient communication

Date: Wednesday 29th May 2019  12-14 pm

Venue: GW 154 

 

 Words that are more frequent tend to be short, and conversely, words that are less frequent tend to be longer, containing more segments. This is pattern makes language more efficent by balancing speaker effort versus the amount of information provided to the listener. But not all segments are equally informative: listeners identify words from the speech stream incrementally, continually updating their lexical search as the phonetic signal unfolds. As a consequence, segments earlier in words tend to contribute more disambiguating information to lexical access than later segments, depending on what similar lexical competitors exist. We expect then that languages should not only optimize the total number of segments in words, but also how informative those segments are in disambiguating from other words in the lexicon. Here I'll show data from a range of languages that this is the case: words that are on average less frequent have relatively more informative early segments, while preserving a 'long tail' of more redundant later segments. 

Second, I'll review our recent work suggesting that this asymmetry in segment information distribution acros the word may influence the evolution of phonological rules which impact lexical identification. In a typologically-balanced sample of 50 languages, we find that phonological rules which neutralize lexical distinctions (e.g., word-final obstruent devoicing in German) are common at word-ends, but very rare at word-beginnings. 

Both of these patterns are consistent with the hypothesis that language change is shaped by a tendency for speakers to reduce redundant phonetic material while preserving more informative material.

 

_________________________________________________________________

 

 

 Mr. Marc Tanti, University of Malta

Title: Making Computers Describe Images Automatically

Date: Friday 21th June 12-14

Venue: GW214

 

Computer programs usually require a programmer to say how to perform a task using step-by-step instructions. Artificial neural networks, on the other hand, are programs that can learn to perform a task just by showing them examples of what they are supposed to do. These are used to do all sorts of things from recognising what object is in a photo to translating the language of a sentence. In this talk, I will be showing how these interesting programs can be trained in order to make them generate text and, more specifically, how to generate a description of a photo. I will be explaining in simple terms how I performed experiments to analyse the different ways of how this can be done and give the advantages and disadvantages of these methods.

 

_________________________________________________________________________ 

 

Calendar
Notices
Linguistics Circle (26 June)

Dr. Olga Bogolyubova, University of Malta

Title: Personality Traits and Subjective Wellbeing in the Language of Social Media

Date: Wednesday 26th June 12-14 

Venue: GW214

Timetables

Class timetables are now available from this page.

For study-units LIN1063, LIN1065, LIN2013 and LIN5063, please click on this page to check the Academic English timetable.

Stress Management Sessions for students

Sedqa, in collaboration with the University of Malta, will be holding two sessions on stress management for students.

Tobacco cessation
Visit the underneath link for more details: 
 
 
Last Updated: 25 June 2019

Log In back to UoM Homepage