Note 1 - 'Dual-use' normally refers to the (military - civil) pair. Here I use the term with reference to the pair (language industry - education industry).

Note 2 - In the CommonSpaces blog, posts should usually be written for everyone (in English), to be eventually translated into other languages. In this case, I preferred to start with Italian, as it is a proposal initially aimed at national audience.

The importance of the mother tongue

Mastering the mother tongue (L1) is an important requirement for foreign language learning and as a basis for disciplinary learning. Ensuring a high level of competence in L1 is therefore one of the investments with the greatest returns for both children and adults

This position does not overlook the fact that, on pain of socio-cultural exclusion, some individuals are in need of learning an L2 quickly, but it combats an overemphasis on the convenience of doing disciplinary studies directly in L2 (e.g. in English), as well as 'popular myths' such as that to learn an L2 it is better to start as early as possible. 

NLP and basic language education

In my opinion, there is currently no significant spill-over effect on language education activities of the advances, albeit significant and continuous, taking place in the field of automatic language processing (NLP), even when the results of this are formally open.

Outside of English there is no critical mass

AI and NLP research by far favour English, because this means following a smoother path (it is easier to find data and compare results) and ensures greater dissemination of publications. But even when the techniques and tools are quite general, there are usually no multi-language experiments in the field of language education. Even the European institutions and initiatives that are supposed to deal with this (e.g. the ELT - European Language Technologies project cluster) are disappointing, as they only give a few 'scattered' funds to equally 'scattered' projects.

Multi-purpose tools for a variety of interests

In general, language training tools derivable from advances in the NLP field are 'multi-purpose'. For example, a text analysis tool could support:

• the assessment of the compatibility of a school text (level of vocabulary, size, structural complexity) with the receptive comprehension capabilities of the target audience;
• a teacher's evaluation of a learner's essay (syntactic correctness, vocabulary richness, etc.);
• self-assessment by the author, who may wish to improve the style of the text.

This multiplicity of uses should allow researchers and teachers with partially distinct interests to be brought together on the same project.

Pedagogical research and field experimentation

Interesting experiments in language teaching innovation have been made as a result of pedagogical research and guidelines issued by professional networks. Many results have been disseminated through journals, conferences, educational materials. But rarely are the reports and materials produced designed for adaptation and reuse in languages other than the original one.

Probably the most valuable results were developed by passionate teachers, sometimes supported by commendable administrators, within pilot projects with multilingual school populations and/or non-profit organisations. Unfortunately, however, personal passion and sacrifice were, for the most part, not supported by adequate language/information technology and tools, and the results were not disseminated and reused in the literacy of children in other countries or other contexts.

Supply or demand side policies?

National and local policies usually deal with the supply of innovation, distributing more or less effective innovation incentives to companies. Incidentally - the notion of 'innovation' is often too general and laden with connotations that are not always compatible with sustainable development.

National and European policies should instead support the demand for innovation (right Prof. Meo?).
In principle, the Erasmus+ projects could help; but the consortia formation mechanism, which favours the inclusion of idle partners, the fragmentation of funding, the precariousness of the projects themselves, of the staff seconded to them, of the stakeholders, etc., ultimately makes them scarcely credible from this point of view.

A working group

I propose setting up a Working Group (WG) on the application of research and development results in the field of NLP to language training.

The participation of school teachers of all levels is essential.

It is desirable to have the participation of

• foreign language specialists
• links with teachers' associations;
• links with national and international institutions dealing with basic education, multilingualism, inclusion;
• researchers and disseminators in the fields of AI and NLP.

Giovanni Toffoli, who in recent years has collaborated with Sapienza in research projects in the field of education, ensures the usability of the OS tools he has developed with LINK srl; Giovanni's collaboration will preferably be in a individual form; LINK will only be involved if this is useful for the pursuit of the goals of the working group under consideration.