Govorjena slovenščina: A Platform for the Structured Collection of Conversational Speech

Abstract

Conversational spoken language resources remain underrepresented despite their importance for linguistic research and speech technology development. Their collection is constrained by methodological, technical, and legal challenges, particularly in the case of private, everyday interaction. This paper presents Govorjena slovenščina, a CLARIN.SI service developed to support the systematic collection, transcription, and archiving of spoken Slovene. The platform integrates standardized workflows for consent management, metadata collection, data submission, and transcription, and is designed as a reusable research infrastructure rather than a project-specific collection tool. Drawing on insights from citizen-science and crowdsourcing initiatives, the paper discusses key design choices related to participant motivation, data quality, and legal governance, and reports on initial corpora contributions enabled by the platform. The approach demonstrates how a standards-driven, legally grounded infrastructure can support the sustainable development of conversational speech resources, particularly for smaller or under-resourced languages.