Short-term syntax learning and generalization effects

Two language therapy procedures- a structured, repetitive drill approach, and a story-based question and answer approach — were used to teach is and are to four language-impaired children aged 4;4 to 6;3, over a four-week period. Effectiveness was measured by three methods: within-therapy data, structured probe data, and at-home spontaneous probe data. While levels of correct productions were high in therapy, production on spontaneous probes remained low. Struc-tured probe performance varied, but was generally substantially higher than spontaneous probe performance. Results indicate that syntax learning depends on how such learning is measured