Linguistics Circle occasional seminar series 2012/2013
Agata Savary (Université François Rabelais de Tours Blois, France) Action - towards a better understanding of the role of multi-word expressions in parsing
- Date: Tuesday 11th June, 2013 at 16.00 hrs
- Venue: Gateway Building 214
COST IC1207 Action, PARSEME (Parsing and Multi-Word Expressions), aims
at increasing and enhancing the support of the European multilingual
heritage from information and communication technologies. This general
aim is addressed through improving linguistic representativeness,
precision and computational efficiency of Natural Language Processing
(NLP) applications. The Action focuses on the major bottleneck of these
applications: Multi-Word Expressions (MWEs), i.e. sequences of words
with unpredictable properties such as "to count somebody in" or "to take
We will present the scientific program of the
Action, which addresses 23 European languages from a cross-theoretical
and cross-methodological perspective.
Manuel Perea (Universitat de Valencia, Spain) Are root letters compulsory for lexical access in semitic languages? Evidence from masked form-priming in Arabic.
- Date: Thursday 6th June, 2013 at 13.00 hrs
- Venue: Gateway Building 214
Semitic and Indo-European languages differ at a qualitative level?
Recently, it has been claimed that lexical space in Semitic languages is
mainly determined by morphological constraints while in Indo-European
languages the constraints are more orthographic (Frost et al., 2005;
Velan & Frost, 2011). One of the key findings supporting this
statement is the absence of masked form priming in Semitic languages,
although nonsignificant trends are usually the case. Here we examine
whether masked form priming occurs in a Semitic language (Arabic) when
one of the letters from the root is replaced in the prime stimulus by
another letter using a large set of items (180 words). Results showed a
significant masked form priming effect with the lexical decision task in
three experiments (including yes/no, go/no-go, and sandwich priming).
The present data suggest that Semitic and Indo-European differ more at a
quantitative than at a qualitative level.
Aleksandra P. Knapik (Polish Academy of Sciences) On the evolution of contact languages
- Date: Friday 17th May, 2013 at 12.00 hrs
- Venue: Lecture Center Room 119
or pidgin varieties are found all over the world on every continent.
They are developed when two or more languages come into contact and form
a new language. Instead of using their own languages to communicate
people enforce a creation of a new language. Most Creole languages are
based on one or two languages. In Jamaica the African slaves found
themselves in a situation where the only common means of communication
was English, therefore the superstrate for Jamaican Creole is English,
or its variety spoken at the beginning of 17th century (Arends, 1995:
56). Essential words which people could not find an English name for,
such as plants, animals; and activities were taken from a variety of
West African languages. There is a wide variety of external
socio-historical factors that determine the rise of contact languages
define the new varieties and decide on their form and structures. Pidgin
languages arise as a consequence of many social and historical
processes which involve political and economic factors in the creation
of quite new and distinct social situations. William Washabaugh and
Sidney Greenfield suggest that pidgin and “[c]reole languages (…)
developed to provide a world of meaning for those caught in this new
life situation and thus to enable them to adapt to the constrains of
that situation” (Washabaugh and Greenfield 1983: 106). Through the study
of a Creole variety in the Caribbean area the presentation investigates
the social and structural dynamics that lies under the phenomenon of
the formation of a new or restructured language variety. It emphasizes
the importance and interplay of historical perspective, socio-cultural
approach and linguistic analysis in the study
of contact languages.
Arends, Jacques, Peter Muysken, Norval Smith (1995) Pidgins and Creoles: An Introduction
. Amsterdam: John Benjamins.
William, Sidney M. Greenfield (1983) “The Development of Atantic Creole
Languages”. [In:] Ellen Woolford and William Washabaugh (eds.) The Social Context of Creolization
. Anna Arbor: Karoma Publishers: 106-120.
Piotr P. Chruszczewski (Polish Academy of Sciences) Selected research paradigms and research directions in anthropological linguistics and anthropological pragmatics.
- Date: Wednesday 15th April, 2013 at 12.00 hrs
- Venue: Lecture Center Room 119
linguistics, and by default also anthropological pragmatics, grew as
sub-disciplines of both anthropology and linguistics. “The intellectual
basis for anthropological linguistics in the United States derives from
Boas ( 1966), whose interests and concerns led to the
anthropological view of language, which is that language is an integral
part of culture (…)” (Klein 2006: 296). Pragmatics enter the scene,
telling the researcher how to analyse the aforementioned phenomena.
Therefore, anthropological pragmatics would be responsible for equipping
the researcher with tools, for it is language and language-oriented
mechanisms of communication, the study of which provides a much clearer
insight into cultural phenomena which often direct the use of language
representing culture from both the synchronic and the diachronic point
of view. “[O]ne approaches language from an anthropological view, which
includes the uses of language and the uses of silence, as well as the
cultural problems involved in silence and speech” (ibid.).
Boas, Franz ( 1966) “Introduction.” [In:] Franz Boas (ed.) Handbook of American Indian Languages
(reprinted by P. Harder). Lincoln, NE: University of Nebraska Press;
1–79. [Reprinted from Bureau of American Ethnology Bulletin, 1911, 40
(1); 18–3. Washington,
D.C.: Smithsonian Institution].
Klein, H.E. Manelis (2006) “Anthropological Linguistics: Overview.” [In:] Encyclopedia of Language & Linguistics
. Keith Brown (ed.). Amsterdam, Boston: Elsevier; 296–304.
Holger Mitterer (Dept of Cognitive Science, University of Malta) Sound to meaning and sound as meaning
- Date: Friday 26th April, 2013 at 12.00 hrs
- Venue: Lecture Center Room 117
language has been characterized as an arbitrary code for meaning (e.g.
Hockett, 1966). By re-combining a small number of “meaningless” units
(i.e. phonemes), a large number of meanings can be represented (“duality
of patterning”). This classical thinking has recently been questioned
in both linguistics, uncovering so-called ideophones with non-arbitrary
sound meaning-mappings in many languages, and psychology, showing
consistent cross-modal mappings. We tested how well real ideophones from
a variety of languages are decodable for naïve listeners and found
that, if prosodic cues are taken away, sound-meaning mappings crumble.
reinforces the importance of arbitrariness, but also raises the
question how listeners can uncover the smaller meaningless units in
spite of speaker and style-specific differences. I will show that
listeners make use of auditory, phonetic, and phonological processes to
deal with this invariance problem. In the process, I will argue that not
context-invariant phonemes but context-specific allophones are the
crucial units for the listener.
Mike Rosner & Jan Joachimsen (ICT, University of Malta) METANET4U: Supporting language technology for European languages
- Date: Thursday 18th April, 2013 at 12.00 hrs
- Venue: Chemistry Building Room 327
was a European project aiming at supporting language technology for
European languages and multilingualism. It was a project in the META-NET
Network of Excellence, a cluster of projects aiming at fostering the
mission of META (Multilingual European Technology Alliance). Running
from 2011 to 2013, the central objective of the project was to
contribute to the establishment of META-SHARE, a pan-European digital
platform which makes available language resources and services for
speech and language processing and which supports a new generation of
exchange facilities for them. At the University of Malta, the project
was based at the Department of Intelligent Computer Systems, in close
collaboration with the Institute of Linguistics. Several existing
language resources for Maltese were collected, assessed and uploaded to
META-SHARE. Others were newly created and added to the collection. We
will present META-SHARE and how to browse, upload and download language
resources. Additionally, we will discuss some issues connected with
working on the platform, i.e., metadata schemes and intellectual
Adam Ussishkin (University of Arizona) Morphological decomposition in Hebrew spoken word recognition
- Date: Friday 15th March 2013, 12:00
- Venue: Lecture Centre Room 117
some research in the domain of spoken word recognition has focused on
whether morphological decomposition occurs during language processing,
very little attention has been paid to languages with typologically
unusual morphology such as Semitic languages. Related work in the domain
of reading comprehension (e.g., Frost et al. 1997, 1998, 2000 for
Hebrew; Boudelaa and Marslen-Wilson 2001 for Arabic) has revealed that
morphological decomposition occurs in visual word recognition, though
the extent to which these results are mirrored in spoken word
recognition is unclear. Recent work on spoken word recognition in
Maltese (Ussishkin et al. in prep) has shown that while word patterns
play no role in language processing, the consonantal root is central to
any model that aims to capture how the lexicon is organized.
this talk, I report on two experiments designed to test whether similar
results obtain for Hebrew spoken word recognition. In one experiment,
we tested for morphological priming by word pattern, and in the second
experiment we tested for morphological priming by consonantal root. If
in Hebrew spoken word recognition behaves like reading comprehension, we
expect to find that both roots and patterns can cause priming effects,
thereby facilitating lexical access. However, our results show that
Hebrew is more like Maltese when it comes to spoken word recognition:
roots indeed facilitate lexical access, but patterns fail to do so. As a
consequence, models of spoken word recognition for Hebrew must
differentiate between roots and patterns, and visual and spoken word
recognition require differing, modality-specific accounts.
Andy Wedel (University of Arizona) Lexical competition influences phoneme-inventory structure.
- Date: Friday 13th March 2013, 12:00
- Venue: Old Humanities Room 105
languages universally show ‘duality of patterning’, in which lexical
categories are composed from a limited set of contrastive but largely
meaningless elements, often termed phonemes. The long‐standing
functional load hypothesis states that phoneme contrasts that
distinguish fewer lexical categories should be more vulnerable to loss,
while those that distinguish more should be preferentially preserved.
et al. (to appear) provided the first clear evidence supporting this
hypothesis showing that within a diverse dataset, phoneme contrasts that
distinguish fewer minimal pairs were significantly more likely to have
been lost. In this talk new data will be described suggesting that
phoneme chain shifts and phoneme splits – both processes that maintain
lexical contrast – are also driven by lexical competition.
new evidence will be presented showing that these effects are present
in natural speech: within the Buckeye Speech corpus, instances of a
phonetic cue that distinguishes its host-word from a lexical competitor
are significantly enhanced relative to instances of that same cue in
words with no such competitor. These results are strongly consistent
with the larger hypothesis that phoneme contrast is influenced by
cognitive processes which promote lexical contrast (Wedel 2012).
Richard Littauer (University of Malta) Using pixel maps to examine typological relations.
- Date: Friday 8th February 2013, 16:00
- Venue: Lecture Centre Room 119
typological differences and similarities between languages can open up
doors for both theoretical analysis and NLP applications. A novel way
of visualising relationships between languages is presented. This was
originally presented at the Language Visualisation workshop in EACL 2012
with Alexis Palmer and Rory Turnbull. The key feature of the
visualisation is that it brings geographic,
phylogenetic,and linguistic data together into a single image, allowing a
new visual perspective on linguistic typology. The data presented is
extracted from the World Atlas of Language Structures (WALS) (Dryer and
Haspelmath, 2011). After pruning due to low coverage of WALS, the
typological data are filtered by geographical proximity in order to
ascertain areal typological effects. The data are displayed in heat maps
which reflect the strength of similarity between languages for
different linguistic features. Finally, the heat maps are annotated for
language family membership. The images so produced allow a multi-faceted
perspective on the data which, it is hoped, will facilitate the
interpretation of results and perhaps illuminate new areas of research
in linguistic typology. Ideas for future work in this vein are
Phyllisienne Gauci (University of Malta) Teaching pragmatics: An empirical study in a foreign language context.
- Date: Friday 30th November 2012, 16:00
- Venue: Chemistry Building, Room 327
examining the effectiveness of teaching L2 pragmatics have increased in
the past few years, showing a growing interest in the area of
interlanguage pragmatics. Results are very encouraging, and agree that
pragmatics is teachable, and pragmatic instruction outpaces the mere
exposition to the target language. Therefore research in the teaching of
pragmatics has now directed its attention to identifying experimentally
the most effective way of teaching. The majority of experiments in this
area compare the effects of different
types of interventions along the implicit-explicit continuum.
study presents the results of a classroom experiment aimed at comparing
implicit and explicit instruction in the context of L2 Italian
teaching. The targeted pragmatic features are lexical and syntactic
devices used to modify the illocutionary force of requests and
complaints. First a pre-test is administered, consisting of a written
discourse completion task, an oral role-play and a multiple choice
discourse completion task. Then for six weeks the three classes receive
different instructional treatments: in one class the
elements are directly dealt with through explicit teaching, in the
second class through implicit teaching, while the third class receives
no specific pragmatic teaching, as it is the norm in the school attended
by the learners. Soon after the treatment the three
tested again with the same instruments used for the pre-test. Finally a
delayed post-test is administered four months later.
expected, the two groups receiving the treatment outperform the Control
group in the post-tests. However, results show some interesting
differences on the relative effectiveness of the teaching methods
adopted. Explicit instruction appears to be effective in promoting the
acquisition of declarative pragmatic knowledge but not as successful in
the development of procedural knowledge needed for online oral
production. The study thus contributes to the debate on implicit vs.
explicit pragmatic teaching with a crucial methodological issue, namely
the role of the testing instruments.
Albert Gatt (Institute of Linguistics) Snagged by the root in the cosnonant jugnle: On reading and morphology in Maltese.
- Date: Friday 9th November 2012, 17:00
- Venue: Gateway Building, Hall B1
A robust finding in research on word recognition during reading is that words (such as judge) whose letters have been transposed to form non-words (such as jugde)
can be recognised easily. This result has been observed not only for
English, but also for other languages, such as Spanish, Thai and Korean.
languages such as Hebrew and Arabic appear to be somewhat different.
Experimental work using the Rapid Serial Visual Presentation paradigm
(RSVP; Velan and Frost, 2007), in which sentences are presented word by
word very rapidly, suggests that readers in these languages are far less
accurate than, say, their English counterparts in recognising words
whose root consonants have been transposed. One interpretation of these
results is that for readers of these languages, the consonantal root has
a special status which is not present in other languages. One possible
reason for this difference lies in their orthographic system, which
tends to privilege consonants and frequently omits vowels. Another
important factor is the centrality of the root in the morphology of
Maltese presents a particularly interesting
test case because although it is a Semitic language, it uses the Roman
alphabet, putting it on a par with languages such as English and Spanish
where orthography is concerned.
In this talk, I describe an
experiment which replicates the methodology of Velan and Frost (2007) on
Maltese readers. It compares the recognition accuracy of Maltese words
whose root consonants have been transposed, to that of English words
with transposed consonants. Our results suggest that Maltese evinces
characteristics that are completely unlike those usually reported for
Semitic languages. Indeed, Maltese readers’ accuracy in recognising
words with consonant transpositions is on a par with their accuracy in
recognising English words under similar conditions.
attempt to interpret these results in the light of two important
characteristics of Maltese: (a) its orthographic system, which is
distinct from that of other Semitic languages; (b) its hybrid
morphology, in which root-based morphological processes co-exist with
stem-based processes. I will argue that, orthography aside, a crucial
reason for the difference between Maltese on the one hand and Arabic and
Hebrew on the other, is the relatively unproductive nature of its
Semitic morphological component. The talk concludes with some remarks on
the implications of these findings for models of word recognition and
Daniel Hardt (Department of International Language Studies and Computational Linguistics, Copenhagen Business School) A Sentiment analysis in American and Danish film reviews.
- Date: Friday 12th October 2012, 18:00
- Venue: Gateway Building, Hall B1
talk is an attempt to answer the question: Do user populations differ
systematically in the way they express and rate sentiment? Large
collections of Danish and American film reviews are used to investigate
this question. Evidence of important systematic differences is found:
first, positive ratings are far more common in the U.S. data than in the
Danish data. Second, highly positive terms occur far more frequently
in the U.S. data. Finally, Danish reviewers tend to under-rate their own
positive reviews compared to U.S. reviewers. This has potentially
far-reaching implications for the interpretation of user ratings, the
use of which has exploded in recent years.