The Department of Defense has over 40,000 positions requiring foreign language skills. But most of the people in these billets lack the full, required skillset. This has a serious impact on military readiness. There are two main reasons for it. Training is slow, often requiring 6-12 month residency, and most trainees fail to achieve the level of proficiency that military training billets require. After resident training, personnel tend to lose their language skills unless they have opportunities to continually practice them. This paper reports on progress toward a novel solution called ALLEARN (Adaptive Language LEARNing) to address this challenge. The solution comprises these elements: 1) A blended learning approach that combines the expertise of language instructors in guiding and managing learning with adaptive personalized instruction; 2) A highly interactive, virtual training environment in which learners apply and assess their language skills; and 3) Adaptive training technology and architecture that selects training content in a manner that accelerates learning, manages decay, and accounts for individual differences in learning rates. Pilot evaluations are being conducted with military personnel seeking to achieve spoken proficiency in Modern Standard Arabic. To better understand learning needs and current barriers to achieving spoken language proficiency, interviews of a total of 17 instructors and 22 students were conducted at US Army JFK Special Warfare Center and School (USAJFKSWCS) and San Diego State University (SDSU). The interviews highlighted a need for online learning activities that provide realistic opportunities to practice spoken language skills in unscripted settings, immediate meaningful feedback, and automated assessments of communicative competence that reduce the burden on instructors. We have developed a system architecture to address these needs. It is an open architecture for language practice and assessment that collects analytics on learner performance and optimizes learning trajectories.