A key method for studying articulatory planning at different levels of phonological organization is masked-onset priming. In previous work using that paradigm the dependent variable has been acoustic response time (RT). We used electromagnetic articulography to measure articulatory RTs and the articulatory properties of speech gestures in non-word production in a masked-onset priming experiment. Initiation of articulation preceded acoustic response onset by 199 ms, but the acoustic lag varied by up to 63 ms, depending on the phonological structure of the target. Onset priming affected articulatory response latency, but had no effect on gestural duration, inter-gestural coordination, or articulatory velocity. This is consistent with an account of the masked-onset priming effect in which the computation from orthography of an abstract phonological representation of the target is initiated earlier in the primed than in the unprimed condition. We discuss the implications of these findings for models of speech production and the scope of articulatory planning and execution.

在不同的语音组织层次上研究发音规划的一个关键方法是掩蔽启动。在以前使用该范式的工作中，因变量是声学响应时间 (RT)。我们使用电磁关节造影术来测量关节RTs 和非单词生成中语音手势的发音特性在一个掩码启动实验中。发音的启动先于声学响应开始 199 毫秒，但声学滞后最多可变化 63 毫秒，具体取决于目标的语音结构。启动启动影响关节反应延迟，但对手势持续时间、手势间协调或关节速度没有影响。这与掩蔽起始启动效应的解释一致，其中目标的抽象语音表示的正字法计算在启动条件下比在未启动条件下更早启动。我们讨论了这些发现对语音生成模型以及发音规划和执行范围的影响。