Your writing has reached a solid level — you can express yourself clearly and correctly on paper. Now it's time to transfer that accuracy to your speech.
The challenge is that speaking happens in real-time. You can't look up a conjugation while in the middle of a sentence. Well, you can, but it's very clunky. By slowing down your speaking, continuing to write and being aware of frequent mistakes, you will improve your speaking accuracy.
The goal isn't perfection. It's to reduce errors to the point where they rarely interfere with communication and you can catch most of your own mistakes as (or after) you make them.
Speaking Analysis — Record yourself speaking (monologues, conversations, or topic talk) and analyze the recording afterward. Listen for errors in grammar, word choice, and pronunciation. This is one of the most effective ways to improve because it separates the act of speaking from the act of analyzing. Speaking Analysis
Topic Talk — Repeated monologues about a single topic to get repeated practice using vocabulary and grammar structures. Helps target errors and improve more than unstructured monologues. It can also give you good topics to take into lessons. Topic Talk
Grammar Study continues. As you analyze your speaking, you'll discover errors that are different from your writing errors. Some of these may be due to differences (small or large) between the spoken and written language.
Continue writing. Unassisted writing remains valuable — it keeps your written accuracy sharp and often reveals areas where your understanding has deepened through speaking practice.
A balance between study, input, and output: The Pillars of Language Learning
Move to Phase 6 when you:
The use of recording and self-analysis to improve speaking accuracy reflects principles from skill acquisition research. DeKeyser (2015) described how automatic, fluent performance develops through a cycle of producing language, receiving feedback (either external or self-perceived), and adjusting behavior in response. Recording yourself creates the conditions for this cycle by separating the production phase from the analysis phase — during speaking, you focus on communication; during analysis, you can afford the cognitive resources to notice errors and think about why they occurred.
The idea that writing accuracy can transfer to speech is supported by Sletova (2023), who found that learners who practiced written recall showed measurable improvements in their subsequent speaking accuracy — a result not seen when learners practiced only through spoken recall. Writing's slower pace allows learners to notice gaps, consult explicit knowledge, and correct errors, building stronger representations of accurate forms. DeKeyser's framework explains what happens next: through repeated practice, the conscious, rule-based corrections made in writing gradually become faster and more automatic — a process called proceduralization. The combination of slow, deliberate written practice followed by real-time speaking practice is designed to support exactly this transition.