Community Research and Development Information Service - CORDIS

Periodic Report Summary 1 - MEMOLI (Measuring and Modelling Language Interactions)

Large-scale migrations result in different languages and cultures getting in contact with each other. It happened so in the past and continues happening so in the 21st century Europe. Jewish languages, such as Yiddish, Ladino, Judeo-Arabic or Judeo-Italian represent eminent examples of such linguistic and cultural interactions. The aim of this project is to better understand the effect of contact on language change and variation, with a special eye on Jewish languages.

A recent proposal by Enoch O. Aboh suggests approaching languages as “hybrid grammars”, in which features from various sources are constantly recombined. Prototypical examples being creole languages, he claims his proposal also applies to any language. Independently of his work, another recent proposal, the approach of Sarah Bunin Benor toward Jewish languages also develops a comparable concept: “Rather than seeing a given Jewish community’s language as a ‘Jewish language’, I propose to see it as the selective use of a distinctively Jewish linguistic repertoire. [...] Jews in any given time and place make selective use of their distinctive repertoire, in combination with the repertoires used by non-Jews [...]” Hence, contact languages, such as creoles and Jewish languages – but to some extent, all languages of the world – display a recombination of linguistic features drawn from various sources.

While this idea sounds convincing for many experts in the field, formal linguistics still requires elaborate analytic tools to develop and test such a model. The objective of the project is therefore to model this process of feature recombination from several repertoires within the framework of Optimality Theory, a linguistic model that comes with a strong cognitive and computational component.

The first two years of the project has laid the bases. In particular, the following work has been done cumulating in the following results:

1. We developed a novel learning algorithm for Optimality Theory in which the learner (the computer running the algorithm) entertains several mental grammars in parallel. Presented at a conference in 2015, and submitted recently as a journal paper, the approach is more directly related to the case of multilingual language learners in a diglossic situation than earlier approaches to the same (so called “hidden structure”) problem.

2. One case of parallel mental grammars – less self-evident than the multilingual mind – is co-morphologies. So we ran an online experiment to test for the presence of such co-morphologies within Hungarian speakers.

Hungarian has front-back vowel harmony: words choose between different forms of a suffix depending on the vowels in the root. Yet, some words vacillate, that is, they may accept both forms. Our intuition was that native speakers unconsciously maintain some “naïve etymology”, distinguishing between old Hungarian words and recent borrowings. So we carried out an online wug-test with several thousand native speakers filling out a questionnaire. The conclusion was not as significant as previously hoped for, but in a conference paper presented in January 2016, as well as in an article to be written up, we argued some evidence might be found for Hungarian native speakers entertaining parallel co-morphologies for words of different origin, which in turn would be an experimental support for the computational models.

3. Another experiment was also run to test whether the behaviour of these “vacillating stems” depend on speech rate: is the probability of choosing the back allomorph of a suffix increased or decreased, if subjects produce the forms under time pressure? In a conference presentation and journal paper, we argued that fast speech increases the chance of choosing the back suffix. This surprising new result provides additional support for the Simulated Annealing for Optimality Theory Algorithm developed by the fellow in his dissertation, and also to be employed in the current project.

Summing up, the project has born first fruits, mainly in the form of conference presentations and papers submitted or close to submission. Additional contributions to the career of the fellow are the collection of novel data (on Hungarian fast speech) relevant for his previously developed model; as well as his move from theoretical and computational to experimental methodologies, a technique new to him, but becoming progressively more central in linguistics.

At the same time, the fellow offered a number of linguistic courses in the Hebrew studies program of the host institution. Hereby he not only reviewed certain topics and literature for his own project, but also contributed to transfer the knowledge to students in a new institution and in a new discipline: from the Netherlands and the United States to Hungary, and from theoretical and computational linguistics to Hebrew studies.

Finally, by taking over administrative tasks and participating in curriculum development, he worked towards better integrating his career into the host institution on a longer term.

Reported by



Life Sciences
Follow us on: RSS Facebook Twitter YouTube Managed by the EU Publications Office Top